I want to check VEP and annovar overlap that is how many variants are annotated same by both of these tools. But I don't how to do it as VEP output contain many transcripts and so there is difference in number of annotations done by both these tools. So what do you suggest I should perform the analysis?
You will not get much help here, I believe. This is an analysis in which you will have to make decisions at each and every step. You cannot reasonably come here and ask us questions at each step, in this regard. The potential issues are endless. I would encourage you to at least annotate from the same sources in both cases: ANNOVAR's main source is RefSeq; whereas, VEP's is obviously Ensembl. As you may or may not know, genes are curated by different groups, e.g., Ensembl, UCSC, RefSeq, HUGO / HGNC, VEGA, etc. You could also configure both tools to only output the canonical transcript - if both disagree on what is the canonical, then just go by the longest transcript.
What did I do when I first ran VEP years ago (c.2014)? - I looked through the manual (online) and decided which parameters to use or not. I did the same for ANNOVAR. In doing this, you will have a greater understanding of how best to cross compare the results of these tools.