Strong tracking Unscented Kalman filter algorithm is an estimation method of approximating nonlinear distribution by the Unscented strategy and tracking strongly system status mutation. The algorithm combines the advantage of strong tracking algorithm robustness, high accuracy of Unscented transform and simple realization. In allusion to the photovoltaic system maximum power point misjudgment problem under partially shaded conditions, a new MPPT control strategy is proposed by combining fixed voltage method with strong tracking Unscented Kalman filter (STUKF) algorithm. The voltage range of the maximum power point is determined by the fixed voltage method. In this voltage range, the small-step control voltage is adopted, the instantaneous power is taken as the estimative value, the power maximum power point is accurately estimated by the strong tracking Unscented Kalman filer, and its corresponding control voltage is obtained. The MPPT tracking accuracy is improved on the basis of assuring the MPPT tracking speed. The mutation status of power is detected by tracking the state, and the MPPT misjudgment problem is avoided. The correctness and effectiveness of the strategy is verified by simulation and experiment.