:Transient stability preventive control of power system can be expressed mathematically as a nonlinear dynamic programming problem with differential equation，and it is very important to develop an optimization algorithm with strong optimization ability and small computational cost． An optimization algorithm mixed with intelligent surrogate model and improved differential evolution algorithm is proposed to solve transient stability preventive control problem． Based on differential evolution algorithm，the extended mutation operation and the parameter adaptive adjustment strategy are introduced to improve the optimization ability of differential evolution algorithm． Meanwhile，in order to improve the speed of solving the optimal power flow， the power flow characteristic and ensemble extreme learning machine regression model are employed to predict the transient stability boundary in the iterative process． The simulation results show that the proposed method not only improves the optimization ability of the traditional intelligent optimization algorithm，but also greatly reduces the solving time complexity．