If it's a new record is because it hasn't been found in a country, region or continent, and this species has been already descripted; but if you are trying to find out if your species is a new species you have to: compare with descriptions in papers which define the genus, interact with specislists of that field, and research in the internet, electronic and phisical journals, etc. Those are the ways to find out the answer you want

I also recommend that you check with local members of the community, often farmers and hunters, who may be familiar with an organism even if scientists are not.  And as Miguel has suggested above, if the finding is within the range of the species, it may just have not been surveyed that well, but specialists should know.

Thanks a lot for your suggestions. I will surely follow your suggestions.

Dear Mr Goyal,

To confirm the discovery of a new species, as mentioned in the previous comments, you will need to be sure that you have verified everything, including books, consultation in herbarium, local people etc.

Now if you are not able to find the closest relatives to classify it, DNA barcoding remains an option, though it is quite costly. Good Luck.

Dear Mr. Luckhun,

Thanks. I also thought of going for molecular analysis (barcoding) to confirm the species.

@ If the plant in question is not recorded earlier by anyone, that can be placed in the category of new record for that region.

I am agreeing with Mr. Miguel and Mr. Ravi Luckhun. You may consider leaf anatomical characters or pollen of closely related species. In identification we have to separate species based on floral/reproductive characters. In preparation of key dont give importance to external vegetative characters, leaf pubescence, petiole length or flower colour which may vary from area to area. Regarding DNA barcoding kindly refer to opinion article on "The seven deadly sins of DN A barcoding" published in Molecular Ecology Resources: 2012. http://web.natur.cuni.cz/~muncling/barcoding2.pdf

Sir I agree with you that for identification we have to depend on floral/reproductive characters for identification. But if the plant is a bamboo species and flowering is a rare event and we have to rely on only morphological features to identify it. How can we go for it?

First ask yourself why do you want to separate your species. Is it so that you can classify them in your own mind? Is it so that others can identify the species in the field to help them to plant, cultivate and make best use of them? Is it so that you can gain an insight into how they have evolved? Is it so that you can conserve them for future generations? This will help you to decide what kind of species concept you should apply.

If you are most interested in floristics and records then you may wish to concentrate firstly on a traditional morpho-geographical approach. This will be necessary if you wish to describe a new taxon, as definitions require morphology. Molecular information alone will not allow you to publish botanical names, it will only provide supporting evidence.

Look closely at the characters currently used to separate known species, understand them and how much they vary within and between species. Then go back to a wider range of plants, preferably in the field, and scrutinise them diligently.

You are right to ask for advice as it is easier to find help in this digital age, although expert knowledge of plant groups is dwindling as the global pool of taxonomists shrinks. Better to get advice at the start than to wait until a reviewer tears your carefully crafted manuscript to shreds.

Consult with Local Herbarium and later KEW is the easiest way to confirm

I appreciate Dr. Chris Stapleton and Dr. Subburaman Senthilumar for the excellent and very useful suggestions.

Superb answer by Chris Stapleton. Splitters always want to split thing until the last piece. Lumpers always want to lump things till the whole will be one.

Any University plant taxonomist should be able to describe the protocol.If you can determine the specimen's family (or a few potential families) it is fairly standard to first contact someone acknowledged to be an expert in that taxonomic family, and then to  send a properly preserved sample.

You will need to first make inquiries within your country- then additional resources may also be used. For example: the Museum of Natural History in New York City... try - http://www.amnh.org/our-research/center-for-biodiversity-conservation

Arbind:

You start by identifying the specimen at least at the family level and search the botanical literature, which taxonomic treatments there are. If you find few works and it does not fill your expectations, you could contact the specialist. If there is not specalist, you could  study the group and the specimens of the herbaria, the local or others where you are sure you can find materials.

Thanks you all for your valuable suggestions.

identifying it to the genus level (with flowers and fruits) and then review the types of species of the genus confined within the geographical area where you found it. If it is not it could be a new species or a new location. Then enlarge the area to rule it a new location

You should be able to identify the family and the genus with general plant taxonomy textbooks and some secondary literature. Once you have established the identity of the genus, you get a list of species (IPNI) and compare the species with the materials you have. If there is no match, you might assume that your material represents a species new to science, but before you publish, you should get at least one more opinion of a trained plant taxonomist.

If, by following the steps outlined by Philip Gerard, you have established that your plant is an existing species, and you can't find it in the latest checklist, Flora and other floristic studies of the area, it can be reported as a new record  

Another aspect that should be considered is the possibility that the plant may be a neophyte. In this case you should first consult books on non-native flora of the area as well as alien plants databases, such as Daisie:

http://www.europe-aliens.org/default.do

When you have any idea about what taxa it is, you could contact with local herbarium of the source region in order to confirm the right determination.

Obviously it depends on the region and the characteristics of the study area. Eg., It's not the same the Amazonian rainforest (where it's reasonable to find new species) that an area near large cities. (enabling environment for the entry of alien plants).

If you have clear which genus it is, you must check the taxonomical keys of the group, compare with herbarium material and review literature. You should try to make it fit in the taxonomical key of the group, if there are morpho characters that differ from its closest relative then you should ask yourself if its a variety, subspp or indeed a new species, then the next procedure of describing starts

In my experience, if you a thorough with the flora of the study area, something new will strike you at the first sight itself.  Then the process is a survey of literature pertaining to the area and  the specimens of the area deposited in Herbaria. If there is no collection of this specimen now a wider survey for the plant is required and also you may approach experts of the family or genus elsewhere.  This way you may be able to find out the species if it already exists or otherwise you may get a chance to publish it as a new species. 

Dr. Alina Freire-Fierro Said that

Steps: 1. Look for any monograph/revision for the entire genus that contains your species. If there is one, follow the identification key. If your plant does not match any of the species described in the key, it is very likely that your species is new. Search also for the specialists on the group to where your species belong. 2. Now, if the revision has been published long ago (e.g.