The combination of PSO algorithm and linear regression formula has proven to help the prototype in producing the correct output of baseline feature for each of the signature. The PSO algorithm generates 90% accuracy in detecting baseline particle (coordinate) but the PSO might produce inaccurate baseline particle if there are too many particles with best position for baseline feature. PSO has the tendency to choose the most optimal solution for detecting baseline particle, but aside from that it is proven to be very useful in detecting baseline feature. The results showed that PSO indeed provided good results on verifying signatures. This was proven by comparing the signatures that come from the same person and those from a different person.