Depends on your instrument and your tuning. You provided us with exactly zero information about both, which allows us to provide you with exactly zero information back.
What kind of instrument do you use? Accuracy of our instrument seems to be definitely much better than 0.5 Da that you used for the search. You can go to 0.020 Da for fragments and I'm pretty sure that 20 ppm for precursors would be fine. BTW, is the intense peak at around 730 m/z in your MS/MS the precursor? If so, increase the collision energy.
Well if you have calibrated ESI-QTOF, first of all, repeat the search with peptide tolerance set to 100 ppm and MS/MS tolerance set to 0.05 Da. With this narrower tolerancy you shouldn't get any outlier that would distort your results regarding overall accuracy.