The track interruption problem that appears in the space-target tracking process will seriously disturb the track fusion and increase the loads of the sensors.Therefore, the study on the spatial target intermittent track correlation algorithm is of great significance.In the traditional intermittent track correlation algorithm, the track obtained by linear extrapolation prediction has big error, which may lead to inaccurate association. Therefore, a spatial target intermittent track correlation algorithm based on trajectory prediction is proposed in the paper. The algorithm tracks the space target based on the dynamic equation, and makes curve-fitting to the filter update value before the interruption to obtain the predicted initial value point. The trajectory prediction is then combined with the target dynamic model. The predicted data is correlated and matched with the state update values of the first few moments after the interruption, to achieve the correlation fusion of the track before and after the interruption. Finally, the simulation results verify the effectiveness of the proposed algorithm.