The accuracy of the synchronous generator model has great impact on power system stability study，and how to quickly and accurately determine the parameters of the generators in a power system is one of the most important issues in the field of power system modeling.In this paper，an online system-wide identification algorithm is proposed with the wide area measuring device and a complete mathematical model is developed for online identification multiple generators in a multi-machine system at the same time.And the multi-scenario parameter identification is proposed to improve the recognition accuracy to cope with the phasor measurement unit（PMU）measurement error.Multi-generator identification is a class of dynamic optimization problems，which is very time-consuming when solving large systems.A reduced-space interior point method is proposed to solve the parameter identification problem with small degrees of freedom，greatly reducing the computational time.Further，the scenario-decoupling strategy is proposed to solve the multi-scenario parameter identification in parallel.Two test cases show that the algorithm under discussion has higher accuracy and can quickly find the identified parameters compared with the traditional stand-alone identification algorithm.