Abstract: Aim at the issue of track before detect based on dynamic programming (DP-TBD) has low detection and tracking maneuvering targets in low signal-to-noise ratio(SNR), this paper proposes a track before detect based on state-weighted dynamic programming. The algorithm extends states transfer set utilizing the characteristic of target motion correlation between the frames, and then the power accumulation value is weighted according to the different possibilities that the current state comes from the previous states, making the power accumulated along the target trajectory. Theoretical analysis and simulation results show that this approach has been an improvement in the performance of detection and tracking maneuvering targets, comparing with the traditional DP-TBD under lower SNR