Abstract: To address the issue of degraded data fusion performance caused by asynchronous sensors in multi-sensor electro-optical tracking systems, a method combining Kalman filter with Lagrange interpolation was adopted to study the preprocessing methods for asynchronous data fusion, achieving time registration of multi-sensor sampling sequences. Specifically, Kalman filter was used to predict sensor data, which suppressed process noise and measurement noise while compensating for sensor delay. Based on the new sampling sequence obtained from Kalman filter, the Lagrange interpolation was then used to normalize the asynchronous data to a unified sampling frequency, which prepared the synchronized raw data for high-performance asynchronous data fusion. A simulation of tracking a tangentially flying target with a multi-sensor electro-optical tracking system under typical conditions was conducted, verifying the effectiveness of the proposed method and obtaining registration data that could reflect the real target trajectory. The results showed that the radar registration accuracy was better than 1° and the electro-optical sensor achieved registration accuracy at the angular level, with a simple and reliable implementation process. The findings meet the requirements for engineering applications in electro-optical tracking systems.