This paper describes a new approach to obtain the global optimization problem of nonlinear mixture chopped stochastic program model. The study focused on the issue of two-stage stochastic with the lack of nonlinearity, which is contained in the objective function and constraints. Variables in the first stage is worth a count, while the variable in the second stage is a mixture of chopped and continuous. Issues formulated by scenario-based representation. The approach used to complete the large scale nonlinear mix chopped program lifting unfounded variable value of the limit, forcing a variable-value basis chopped. Problems reduced is processed at the time of chopped variables held constant, and the changes made during discrete steps, in order to obtain a global optimal solution.