do I get right PCR product - (Dec/29/2007 )

hmm... it looks to me that there are two bands in the mix. Alternative splicing perhaps?

I agree, there are multiple bands. Your template may actually have different sizes (perhaps alternative splicing), but I would first try raising the annealing temperature to increase specificity. You could also be amplifying from genomic DNA along with the cDNA -- the gDNA would not be spliced. Have you done a no reverse transcriptase negative control? What is the expected size of the unspliced gene? Can you do nested or semi-nested PCR? You may have to redesign the primers.

Changing from 57 to 58 is unlikely to do anything. Change to 60, 62, 64, or do a gradient. Are your primer sequences handy? It would be good to check them. Cut back to 35 cycles.

I did PCR again and cut the number of cycle from 50 to 30
this is my result
do I get the right product now?
my gene of interest is 378bp[attachment=4015:upload.jpg]

Looks a lot better than 50 cycles.
To be sure, you could sequence your PCR product or do a restriction digest (in case there's a suitable restriction site in your amplicon) and check whether you get the correct fragments by gel electrophoresis.

That looks a lot better, but the gel lanes are overloaded. Dilute your sample 10x and rerun the gel to get better resolution. You could also run the gel longer to improve resolution where you care. It does look as if you have some primer-dimers being formed at low MW, but these may not matter, depending on what you are doing next.

Incidentally, the black region at the bottom indicates that you are not adding dye to the running buffer, and your EtBr is running toward the negative electrode. You can fix this by adding EtBr to the positive terminal running buffer, or by post-staining.