Abstract: Considering the issues arising from the cell averaging constant false alarm rate (CA-CFAR) algorithms, such as target masking in multi-target scenarios, a modified CA-CFAR algorithm based on ratio (RCA-CFAR) is proposed. On this basis, the maximum number of tolerable interference targets for the RCA-CFAR algorithm is theoretically derived, and the relationship between the maximum tolerable interference targets and the number of reference cells is established and proven. Furthermore, the RCA-CFAR algorithm is extended to the 2D RCA-CFAR algorithm by integrating it with the CA-CFAR algorithm. Through a staged processing approach, the RCA-CFAR algorithm is applied for pre-detection in the range dimension, while the CA-CFAR algorithm is utilized for target detection in the Doppler dimension. By this method, the detection threshold in regions with neighboring multi-targets is reduced, and detection efficiency is improved. The effectiveness, practicality, and superiority of the proposed algorithm are validated through simulation and experimental results.