The flexibility of distributed multi energy technology promotes the development of smart energy stations. However, due to the nonlinearity of the system and the complexity of modeling, the constraints related to the uncertainty of power, thermal network and energy demand are often ignored in the optimization problem of smart energy station. Therefore, a smart energy station operation optimization strategy considering system uncertainty is proposed. The strategy takes the energy expenditure cost as the objective function and uses iterative modeling, including mixed integer linear programming (MILP) and linear approximation of nonlinear energy network equations. In the MILP optimization stage, the operation plans of controllable equipment in all energy networks are modeled considering the uncertainty and the linear approximation of the integrated network. Then the nonlinear comprehensive energy network model is established, and the accuracy of the linear model is improved through the iteration between the two models. The multi-dimensional linked list is used to effectively model the uncertainty and improve the calculation efficiency. The proposed strategy is tested in the simulation. After using the strategy for optimization, the energy expenditure of the test system is reduced by about 3%, which verifies the effectiveness of the strategy.