To tackle the problem of single measurement and low precision of GNSS-R sea surface wind speed inversion method, a segmentation Doppler waveform integral matching inversion algorithm based on Delay Doppler Map (DDM) is proposed. The method utilizes a new DDM observation parameter and, after training it can invert the sea surface wind speed. The algorithm is used to establish a wind field inversion model based on the given conditions, using CYGNSS data as the simulation input. The simulation results show that the root mean square error of the inversion wind speed is 2.01 m/s at the wind speed of sea surface of 20 m/s, which is better than the peak power method and leading edge slope (LES) matching method used in the traditional GNSS-R wind speed inversion.