In order to meet the requirements of accurate estimation of traction motor stator current for train traction drive system in the fields of traction motor condition monitoring, anomaly diagnosis and prediction, a real-time estimation method of traction motor current is proposed based on multi-condition least squares support vector machine (LS-SVM) model. Combined with traction motor parameters and the test measurement data of torque and speed, the stator current is estimated by using the traction motor mechanism model. Then, according to the train operation law, multiple operating conditions are considered. Based on the relevant variables of historical operation data, the LS-SVM estimation model of stator current under each working condition is established, and the influence of different operating conditions of traction motor on the accuracy of the model is studied, so as to realize the effective real-time estimation of traction motor current under all working conditions based on the multi-condition model. Finally, the effectiveness and feasibility of the proposed method are verified by the actual operation data.