Abstract: With the in-depth research, the amount of simulation calculation explosiving increase and improve the computationalefficiency is particularly important. Analysis of the Monte Carlo method (MC) and the difference method of successive iteration(SOR) of the advantages and disadvantages, this paper presents a new type of hybrid algorithm SOR-MC, the use of the Monte Carlomethod random walk construct an SOR initial iteration value, and improve the solution speed. And the use of variable grid to furtherimproves the accuracy of SOR-MC algorithm. Through multiple numerical experiments verify that the hybrid algorithm and itsimprovement can significantly reduce the number of iterations, thus effectively enhancing the computational efficiency.