Abstract: ‍ ‍In dense clutter environment， the traditional signal processing methods and radiation resource optimization algorithms are hard to be applied when target detection and tracking are regarded as two independent stages. By using the closed-loop structure of detection and tracking integration， the detection threshold and radio frequency radiation resource can be reasonably optimized to further improve the radio frequency stealth performance. Therefore， this paper proposes a joint optimization algorithm of detection threshold and dwell time allocation for target detection and tracking in radar network. By integrating the detection and tracking， and introducing the feedback of the tracker information into the Bayes detector to adaptively adjust the detection threshold， the tracking accuracy of the target can be improved. Firstly， the average detection probability in associated gate is derived as the metric for target detection performance. With the information reduction factor， the predicted Bayesian Cramér-Rao Lower Bound is derived as the metric for target tracking performance. Secondly， based on the constraints of certain target detection and tracking performance requirements， limited dwell time resource， the joint optimization model of detection threshold and dwell time allocation for target detection and tracking in radar network is established. By optimizing the detection threshold and dwell time of each radar， the total dwell time resource consumption of radar network is minimized to improve its radio frequency stealth performance. Furthermore， the problem is solved by combining sequential quadratic programming algorithm and improved probabilistic data association algorithm. The simulation results show that compared with other algorithms， the proposed algorithm can consume the least dwell time resource to meet the certain target detection and tracking performance requirements and achieve the best radio frequency stealth performance.