Robust optimization is a method to process optimization problem under uncertainty. The current robust optimization methods have some deficiencies in application conditions and probability utilization. Based on the chance constraints programming, two kinds of robust constraints according to two different kinds of probability distribution of the stochastic parameters are proposed, and a novel robust optimization model is proposed. The feasible solutions of this model can be controlled to satisfy the robust index. This model can be used in the situations that both sides of the constraints contain stochastic parameters, and can be easily extended to non-liner models. The simulation results illustrate the validity of the model.