Abstract: The photovoltaic grid connected system model has the characteristics of high dimensional, multiple scales, and non-linearity, which present challenges to grid connected control and stability analysis. This paper proposes a dimensional reduction modeling method for photovoltaic grid connected systems based on singular perturbation theory considering multiple time scales. At first, a grid connected system model is established based on the photovoltaic output characteristics, and its state space model is obtained by lingering the model. Then by using singular perturbation theory and eigenvalue analysis methods, the photovoltaic small signal model containing multiple time scales is subjected to dimensional reduction processing to obtain the dimensional reduction model. At last, Simulation results show that both the reduced dimensional model and the original model reflect the oscillation mode of the photovoltaic grid connected system, but the reduced dimensional model has a faster recovery time, which verifies the effectiveness of the proposed method.