<p>The uneven load distribution of involute splines is caused by machining errors, which significantly reduces their fatigue life. In this study, a real-time life prediction method based on digital twin technology was proposed for splines with different tooth clearances: the equivalent force chemistry model of splines was established to derive the key tooth load distribution model of splines; Combined with the improved stress correction algorithm, the S-N curve of bending stress of spline tooth root was constructed, and the fatigue life prediction model of spline was established. The Simulink-based digital twin system integrates rainflow counting with real-time torque feedback. The verification case shows that the fatigue life predicted by the above method is more than 10 % smaller than that predicted by the traditional method, which is more in line with the actual situation and provides key technical support for the reliability design of high-precision splines.</p>