Ashish,

first, you need very good collection of accessions (about 30) identified by other methods as  your target plant species and all (at least most) of the related species (the number depends on the genus). GenBank accessions are sometimes misidentified and your task to confirm the names by morphological and cytological analysis.  Having DNA of all your samples, you can make bulks (up to 10 individual DNAs) and look for best method for DNA fingerprinting: AFLP is the most repeatable and informative, ISSR has less intra-species variation, RAPD is the most simple and cheap, and find optimal PCR conditions for the best repeatability.  Use the most strict PCR conditions for repeatable amplification and two-primer system for ISSR or RAPD for larger informative band number. Make PCR fingerprints for each primer combination at the same day, at the same PCR machine, with the same PCR mix. Do not change the instrument and PCR kit /stock during the whole experiment.

When you find specific band for bulks of your target species, you must  repeat the analysis with individual accessions, using at least 3 different plants for each one. Confirming specificity of the fragment, you should clone it (the best way) or sequence directly, if the fragment is amplified alone or with a few distinct bands, and re-amplification of the purified target fragment gave one  major band only. After sequencing you can design several candidates for SCAR primers using if possible homologous sequences from GeneBank.  Test them on all your plants again. Optimize PCR conditions to obtain best results (only one band and lowest cycle number). 

Thanks sir provide me Valuable Information.

sorry for mistyping : ISSR has less intra-species variation (not Inter-species)