Multi-static radar systems are capable of detecting and tracking stealthy targets. However, the state-of-the-art detection and tracking algorithms in a centralized fusion framework suffers from high computational complexity. Therefore, an approach, namely distributed compressed sensing based joint detection and tracking, is proposed for the multi-static radar system, which reduces the computational load largely, in a centralized fusion framework. In the proposed approach, a distributed compact sensing matrix pursuit(DCSMP) algorithm is firstly adopted to reconstruct the sparse grid reflection vector by using distributed compressed sensing matrix pursuit algorithm. The outputs of the DCSMP algorithm are directly fed as instantaneous measurements to the track-before-detect(TBD) tracker, which removes the false measurements and correctly associates the target-generated measurements to the respective targets. Numerical experiments are given to illustrate the correctness of the proposed algorithm.