Abstract: The steady state model of the heat pipe heat-exchange system is built using the mechanism analysis approach.For the unknown key parameter estimation in modeling,the particle swarm optimization algorithm and recursion algorithm with restricted memory are used based on the actual running data.Because of the change of model parameter,a RBF neural network is used to build a quantitative relationship between input variable and model parameter,and the validity of this method is verified using field data.Finally,a precise steady-state model is obtained.It lays a solid foundation for further system analysis and real time optimization.