Arc fault is an important cause of electric vehicle fire. The complex driving conditions of electric vehicles, the high voltage and current in the motor and its drive system, and the great randomness and concealment of the arc fault, make it difficult to conduct the real vehicle experiment. Therefore, a method using a fractional horsepower motor and load system is proposed in order to simulate faults and carry out a large number of experiments to research arc fault characteristics. Firstly, based on the load torque calculation and equivalent scaling, a simulation experiment platform is built to collect the line series arc fault current of a three-phase permanent magnet synchronous motor. Secondly, a driving motor and load simulation system of electric vehicle controlled by space vector pulse width modulation is built via MATLAB software, and through introducing and improving Cassie arc fault models, the line series arc fault of the electric vehicle threephase permanent magnet synchronous motor is simulated and analyzed. Finally, a feature extraction method based on the proportion of flat shoulder width and the proportion of wavelet packet decomposition energy is used to quantitatively evaluate the results by comparing the simulation and the measured data. The results show that the relative average error of the proposed Gaussian arc fault composite model is only 7.6%, the smallest one among all models. The constructed simulation system can effectively simulate the actual line arc fault, which is of great significance for the prevention and control of electric fire in electric vehicles.