Abstract: In order to obtain the robust detection capabilities of high-speed and high maneuvering weak targets, in this paper a fast hybrid integration detection algorithm is proposed based on dynamic programming(DP). The algorithm adopts Markov motion model, allowing targets to maneuver in any form within a given acceleration range. The coherent integration within the segment is realized through Keystone algorithm, then the non-coherent integration between segments is realized through DP. In the implementation process of DP, the original algorithm is improved, and the amount of computation is kept within a reasonable range. Comparing the proposed algorithm with other typical algorithms through radar echo simulation data, in terms of detection performances the proposed algorithm has obvious advantages over other commonly used coherent, non-coherent and hybrid integration detection algorithms for high-speed and high maneuvering targets, and is only slightly weaker than the Radon-Fourier transform-DP-binary integration algorithm based on strict DP solution. In terms of computational complexities, the proposed algorithm has obvious advantages, and it outperforms coherent integration and hybrid integration algorithms when the integration time is long. The test results show that the proposed method achieves an effective balance between detection performance and computational complexity, and can be well applied to high-speed and high maneuvering weak target detection.