Abstract: In the tracking task of distributed radar networks, low signal-to-noise ratio will lead to detection probability of the target less than 1, which may lead to the interruption of target track. This paper proposes a power allocation algorithm for distributed radar networks to track stably for a long track. The track continuity and the tracking performance of the system are both guaranteed by optimizing the power allocation for radar nodes. First, signal and measurement models of distributed radar are established. Then, the Bayesian Cramér-Rao lower bound (BCRLB) under uncertainty measurements is derived, and a mathematical model of the power allocation problem is established. To efficiently solve the optimization problem which contains complex non-convex constraints, a self-constrained power allocation (SCPA) algorithm based on convex optimization is proposed. The simulation results show that the proposed SCPA algorithm can ensure that all target tracks are not interrupted in the whole tracking process, while keeping good tracking performance.