Abstract: For the problem of real-time orbit determination problems of spacecraft with limited communication and computation resources, an orbit determination algorithm of unscented Kalman filter （ELUKF）spacecraft based on event-triggered learning is proposed. Based on the unscented Kalman filter method, an adaptive learning algorithm driven by a stochastic event-triggering mechanism is designed to achieve real-time orbit determination under dynamic learning frequency of finite system. The proposed method can efficiently reduce the bidirectionalm communication times between the ground station and the spacecraft while ensuring the orbit determination accuracy, the communication and computation resources of the spacecraft are effectively saved.