In recent years, biometrics has become one of the most promising identity verification technologies. For the limitations, it is difficult for single mode biometrics to meet requirements of modern identity verification. The paper introduced several common biometrics verification methods and procedures. The limitations of single mode biometrics were also provided and data fusion technology was introduced to solve the problem. On the basis of this, Particle Swarm Optimization (PSO) neural network algorithm was used to construct multi biometrics verification system. The results of experiment based on the method shows that it can achieve better identity verification result and meet requirements of practical applications.