Aiming at the problem of large extrinsic calibration error of micro electro mechanical system (MEMS) Lidar and camera, an extrinsic calibration method based on key points of calibration board is proposed. First, the multi-frame point clouds are superimposed for preprocessing, and then the edge line of the calibration board is fitted based on Hough transform to determine the key points. Finally, the constraints corresponding to the key points and normal vectors are set up, and the optimized calibration external parameters are obtained by using the point and plane correspondence algorithm. The experimental results show that the proposed method can accurately extract the key points, make the average error of calibrated external parameters lower than the existing external parameter calibration methods, and improve the accuracy of calibrated external parameters.