The invention discloses a differential protection fault detection method, a differential protection fault detection device and a differential protection fault detection system of an energy storage tramcar super capacitor charging device based on a BP neural network. The method comprises the following steps: collecting real-time voltage and current data at the input side of the charging device, real-time voltage and current data at the output side of the charging device, voltage and current fault data at the input side of the charging device and voltage and current fault data at the output sideof the charging device; calculating to obtain input-output current differential data based on the collected real-time data; taking the collected real-time data and current differential data as inputlayer neuron data of a pre-built BP neural network model to obtain output data of the BP neural network model; and judging the positions of fault points according to the output data of the BP neural network model. The detection accuracy is high, the overhaul time can be shortened, and the troubleshooting work efficiency of the fault points is improved.