Summary: According to the international price of crude oil and its change of recent data, we give the matrix of the amount of international crude oil prices change state transition probability (or frequency). According to the international crude oil price forecasting error minimum expectation and variance as the optimal target, taking the international crude oil price forecasting error minimum expectation and variance as the optimal index, a double random integer programming model to predict the price of international crude oil is proposed. Then we discuss the existence of an optimal solution. According to the constraint characteristics, an optimization algorithm is constructed. At the same time, according to the current domestic refined oil pricing mechanism, we predict the domestic refined oil price adjustment by applying the optimization algorithm which is proposed in this paper. The empirical analysis shows that the model in this paper is of certain accuracy and practicability of the optimization algorithm.