Power system monitoring and control relies on the result of dynamic state estimation. Installation of PMU in power grids in recent years makes it possible to study the dynamic properties of power system. However it's hard to replace PMU on all buses with the conventional measurement of SCADA system in near future and there are lots of traditional measurements of SCADA system. Therefore, it is reasonable to use both measured data. This paper presents a hybrid dynamic state estimation algorithm by the PMU and SCADA measurements and because we have different measurements gained by PMU and SCADA system with different sampling rates, we can apply data fusion methods to improve the estimation results. The proposed method is examined on 14-IEEE buses power systems and the results show estimation improvement by the hybrid approach.