Abstract: In view of a variety of wireless access technology selection in the integrated environments of heterogeneous networks, an improved PSO-FNN-based vertical handoff algorithm is proposed. The algorithm consists of four modules including forward difference prediction of received signal strength, pre-decision, fuzzy neural network processing and the decision of network selection. And it is combined with the forward difference prediction for the signal strength of next time to improve the handoff accuracy. By applying the pre-decision module, the number of samples entering into fuzzy neural network is decreased. Received signal strength, network bandwidth, network load, terminal moving speed are taken into account as the decision factors in the fuzzy neural network processing module. And using the improved PSO which shows better search ability to optimize FNN. The simulations show that the proposed algorithm can select the network which has better performance in terms of service response time and throughout and can improve customer service quality.