Abstract: This paper proposes a method which combines extreme gradient boosting（XGBoost） with particle swarm optimization algorithm of self-adaptive parameter based on ant colony algorithm（ASPSO） to achieve the transient stability preventive control of power systems. First, the mapping relationship between system operation features and transient stability is quickly learned by the XGBoost model,and the importance ranking of features is given to provide certain model interpretability. Second,the trained XGBoost is embedded into the transient stability constrained optimal power flow model as transient stability constraints. Furthermore, the ASPSO algorithm is used for iterative solution to ensure the transient stability of the system while considering the minimization of the generator active output adjustment, and formulate the corresponding preventive control strategy. Finally,a case study on IEEE 39-bus system provided by PSS/E is performed to demonstrate the effectiveness of the proposed method.