Multichannel synthetic aperture radar ground moving target indication has been widely used and get great attention nowadays. Traditional adaptive methods like space-time adaptive processing (STAP) need to calculate and search the maximum of the test statistic pixel by pixel over a large range of radial velocity, which will cause great computational burden. Nonadaptive methods like displaced phase center antenna (DPCA) will cause signal loss of moving targets, thereby degrading the performance of moving targets detection and radial velocity estimation. To improve the problem mentioned above, this article proposes a novel two-step scheme based on greatest of (GO)-DPCA and Local STAP. In step I, GO-DPCA utilizes the complex images of the channels with different baselines for joint moving target detection. Then, in step II, only the local data of moving targets detected in step I needs to be extracted from the original image domain for clutter suppression and radial velocity estimation simultaneously by STAP. Since the processing gain of moving targets depends on the radial velocity and baseline length, the detection performance of moving targets with different radial velocity can be improved by proposed GO-DPCA compared to conventional DPCA. Considering the moving targets usually distribute sparsely in the scene, step II can avoid most computational burden caused by processing pixel by pixel. Finally, the experimental results of both simulated and acquired data are given to validate the effectiveness of the proposed method.