With the rapid socio-economic development, global issues such as the energy crisis and environmental pollution have become increasingly severe. Electric vehicles have garnered growing attention due to their environmental friendliness, simple structure, and high energy efficiency. Electric drive technology is regarded as the most ideal universal driving solution for pure electric vehicles, hybrid electric vehicles, and fuel cell vehicles. The electric drive axle, as a new type of integrated mechatronic transmission system, is being increasingly adopted owing to its advantages such as high drive efficiency, flexible spatial arrangement, and facilitation of digital and automated chassis control. This study focuses on reliability testing methods for the differential in electric drive axles, mainly including the extraction of reliability test conditions and feasibility analysis of the test scheme. The main research contents are as follows: First, based on specific parameters of an electric vehicle, a Simulink model of the motor and differential is established using existing data, and a full Carsim model of a four-wheel-drive electric vehicle is constructed. Working data of the rear drive axle differential are obtained through simulation tests under typical operating conditions. Next, the working data are preprocessed. Principal Component Analysis (PCA) is employed to reduce the dimensionality of the test data and screen for principal components that meet the requirements. The selected principal components are then subjected to K-means clustering analysis. Using the simulation data, reliability test conditions for the differential are constructed. Based on the test process and literature review, shortcomings of existing testing methods are analyzed. Finally, based on the extracted test conditions, an accelerated fatigue test method for the differential is designed, and the feasibility of the test bench is evaluated.