Computer numerical control (CNC) machine tools mainly use transmission elements such as screws and gears as drive systems. To obtain better accuracy, equipment manufacturers apply pre-pressure to the screw to reduce the occurrence of backlash errors. Backlash errors are one of the important factors affecting positioning accuracy. When the pre-pressure fails, a backlash error occurs. Backlash error can cause abnormal vibration or drive delays in CNC drive systems, leading to reduced transmission efficiency and poor positioning accuracy. Therefore, in most research, vibration sensors are installed on the fixed end of the screw or the nut to detect abnormal phenomena and to estimate the backlash error using the algorithmic model. This traditional method has been established on the basis of the relationship between the amount of vibration and backlash error using an algorithmic model. The backlash can be estimated using this traditional method. However, this traditional method cannot directly sense backlash errors, and the accuracy of the measurement is limited by the algorithm model. Therefore, in this study, we attempted to use a photoelectric switch sensor combined with a microcontroller unit to detect the backlash errors of the CNC drive system. This system can detect a time change of 10 µs. After error corrections, the gain values of each backlash amount and delay time can be used to predict the backlash amount. Through the final verification with commercially available instruments, the system accuracy was found to reach 0.001 mm on the linear axis and 0.001° on the rotating axis. Owing to the streamlined sensing principle of the proposed system, in the future, the system can be integrated into the relevant rotary table components or within the CNC drive system.