In order to determine an optimal route, network performance and cost of many network providers must be compared when end-users visit the content provided by content providers under certain QoS constraints in multi-provider network. The correlative network information is collected by mobile Agent. The delay and cost between any two nodes of network is set to random variables. A minimum model with the expectation of cost and delay value is presented in a stochastic network. The optimal solution of the mobile Agent route from a service provider to a content provider is computed by using genetic algorithms. The obtained simulation results show the effectiveness of the above approach.