A new access selection algorithm based on D?S evidence theory is proposed in order to tackle the problem of existing access selection algorithms that fails to consider the network load sufficiently in heterogeneous wireless networks. Firstly, proper utility functions are determined based on the users’ mobility as well as the network performance and fused by using D?S evidence theory to reduce the adverse impact of parameters’ uncertainty and time-varying on access selection. The candidate network with the highest trustworthiness is eventually selected as the optimal access network. Simulation results show that the proposed algorithm can guarantee the Quality of Service(QoS) of different services and effectively decrease the overall call blocking probability by dynamically adapting to the network load conditions while taking into account the subscriber characteristics．