During method development it has been observed that peak shape of analyte is poor/peak splitting is observed when sample solution is prepared in a solvent other than the mobile phase,why is it so. please inform me regarding sample preparation of drugs unstable in solution(ie moblie phase).

Ideally, sample solution and (starting) mobile phase should be matched, deviation from this produces the results you have found. The reason for this is that you are changing the conditions within the column, the bigger the mis-match, the bigger the changes, the larger the effect, in general terms. Also, the larger the volume injected, the more obvious the problem.
My suggestion is to look at changing the mobile phase, it must be stable in some solution or it wouldn't stay in the body in that form for long!
Even extremely labile material can be chromatographed successfully; I separated chromatographically and collected a cis isomer that changed to the trans form at temperatures above -80C, and in the presence of light, oxygen and metallic ions, and the product now sells for more than $3,000,000 a gramme!
I think your problem can be solved,
Cheers
Bryan Wallwork

well, i think the problem is not deu to instability of your compound. i have seen this before. is the gromatography disturbed wih small injection-volumes ( say 2 µl)? if you desolve your samples in a "strong" solvent (say acetonitril) then you inject a plug of strong solvent when you inject your sample. a part of your analyte chromatographs with this plug and a part of your analyte gets retained on the column. this gives a bad chromatography. as a rule of thumb in our lab we inject max 10 µl of sample in a strong solvent. if we want to inject more, then the sample must be desolved in mobile phase or an even weaker solvent. for one gradient-analysis we inject 300 µl of sample in 10 % acetonitril and still get a good gromatography.
so the strenght of the solvent of your sample is an importent thing.