To address the issues of signal interference and the difficulty in extracting weak features of assembly defects in the current computer numerical control (CNC) machine tool feed system assembly quality inspection, this paper proposes an online assembly quality detection method based on motor current signature analysis (MCSA). This method leverages the non-contact measurement, rapid response, and convenient data acquisition characteristics of current signals, offering a new approach to tackle the aforementioned inspection challenges. Firstly, from the perspective of mechanical stress, this article systematically analyzes the relationship between the loads on the machine tool feed system and the required motor torque under three motion states: constant speed, acceleration, and deceleration. Subsequently, the machine tool feed system is established under typical assembly defects such as misalignment between the guide rail and the screw, excessive screw preload, and protective cover jamming, to study the linear mapping relationship between the motor current signal and the applied load. This allows for theoretical analysis of how load variations caused by different assembly issues affect the motor current. Finally, a signal acquisition system for the assembly quality inspection of the machine tool feed system was established. Different typical assembly problem conditions of the feed system were simulated, motor current data under different conditions were collected, and correlation verification experiments as well as simulation experiments of common assembly quality problems of the machine tool feed system is conducted. Relevant experimental data indicate that the motor current is positively correlated with the feed system torque; assembly quality problem simulation experiments show that different assembly defects can have specific effects on the motor current signal, causing specific changes in the features corresponding to the motor current signal. By analyzing the abnormal changes in the machine tool motor current during the operation of the feed system, effective detection of the assembly quality of the CNC machine feed system can be realized, laying the foundation for subsequent research on identifying assembly quality issues based on current signal analysis.