the species code contains 6 letters which come from the first 4 letters of genus and first two letters of species. However although you use 6 letters of species codes, you still have the same code of different identification. In this case I will change the last letter of the code by using the last letter of species.

I agree with Anuttara, except the code could be 2, 3 or 4 letters of the genus name with 2, 3 or 4 letters of the species name, whatever you prefer, as long as you provide that all-important key.

In my own experience, the best solution is to use the whole species name to avoid errors that might arise from using the wrong USDA PLANTS codes. Codes could be used if you only ever have to deal with a few dozen species. But for hundreds, I rather take the extra time to wright enough of the name to know what it is. For the Access database, I use a separate field for genus and specific epithet and allow queries within the fields to provide the complete spelling as I type. Generally, after only two or three keystrokes I can press enter for the correct name. After the genus, the suggested epithets are confined to those of the genus just entered. The lookup names can be made more efficient if you narrow the list of species to those that are known from the state or province that you've inventoried. Queries can take care of fixing synonymy or assigning codes if needed if they already exist (e.g. USDA PLANTS codes). I should add that I do use a concatenated species name as part of the database key unique to a species-plot combination so that I can relate to a table of attributes belonging to individuals of that taxon within the plot (e.g. tree diameters by species by plot). This was done to allow me to cut and paste records from a different copy of the database.

If I had to come up with my own codes, I'd use three letters per word instead to two, to further avoid resorting to breaking ties using letters from other portions of the name, or worse, numbers that have nothing to do with the species name except in context with a comprehensive list. It is better to use more letters for genus than for epithet, however, since there are usually more genera than species within a single genus. So perhaps genus 4 and epithet 2, and a 7th letter only for infrataxa.

I would use full species name. That way everything is clear. If you do revisions to the manuscript (or website) you do not have to translate codes. If there are a large number of species, then you may start to forget the codes. Relearning them will be painful and may result in more mistakes.

I worked on moths. So there were numeric codes already established. Also, since it was a single order, the codes were unique to a species. I could use the codes for analysis, but names in the manuscript.

There is a catalog for Diptera and one for Hymenoptera. I am fairly certain that the catalog for wasps has numbers. I can't remember if "Diptera of North America" has numbers as well. There is a book by Ross Arnett for the Coleoptera of North America.

A problem will arise if you have multiple orders. Species 5683 may be present in each catalog, but would represent a different species. In this case I might try something like C5683 for species 5683 in the Coleoptera. You probably should use the first two letters of the order to avoid Hymenoptera versus Hemiptera problems.

In general, people will be grateful if you do not invent another coding system when a suitable one already exists. If you must make a new coding system, then use as much of the old coding system as you can, or make the smallest change possible. Also, if you can use an existing coding system, then you do not have to provide a key (though it would be helpful to some if you did provide a key). Depending on what group(s) you are working with you might try to find a local expert to ask what catalog is most accepted. It is a really big pain to use one catalog numbering system only to have it "updated" and then having to renumber all your species to conform to the new system.

I completely agree with Thimothy - "people will be grateful if you do not invent another coding system when a suitable one already exists". I think, any coding is not-parsimonious thing and should be used fonly for internal, private usage. Even names of  plant species authors we are writing now in full in accordance with IC for Plant etc nomenclature. To my opinion, only full scientific names should be used to make data available to public.

I agree to comments that universal only one system is exist so better follow it. For personal purpose you can use any manner.

Definitely agree with the full species name use. If you can precise which Flora or Check-list you use as a guideline, it would be even better because sometimes the same name is understood differently depending on the authors.

However, if in your country or area you have a widely used coding system (as we do in France) you can also provide the codes, so that people using that coding system can easily use your data along with other data encoded within that system. That would be the cherry on the cake.