Water is the only atmospheric parameter with three-phase states. The study on distribution and variation in three-phase water is of great scientific significance for understanding cloud microphysics, cloud precipitation physics, and water circulation, especially in the fields of artificial weather process. In the Raman lidar detection technology of three-phase water, it is necessary to solve the problem of high-spectral spectroscopic technique to ensure fine extraction of the echo signal and the detection with high signal-to-noise ratio (SNR). Considering the Raman spectrum characteristics of three-phase water, the influences of filter parameters in the Raman channels on the overlapping characteristics are theoretical simulated and discussed in detail, and the SNR is investigated as well. Regarding the fact that optimal solution can be obtained for neither overlapping nor SNR at the same time, an evaluation function method based on the multi-objective programming problem is proposed to analyze the optimal filter parameters. The results show that the minimum overlapping value and the higher system SNR can be obtained when the central wavelength and bandwidth of the filters are determined to be 397.9 nm and 3.1 nm, 403 nm and 5 nm, 407.6 nm and 0.6 nm in solid water, liquid water and water vapor channel, respectively, and thus the optimal design can be realized for synchronous detection Raman spectroscopic system for three-phase water. Further simulation results show that effective detection can reach above 3.6 km in the daytime and over 4 km on sunny days under a system factor of 1800 J·mm·min for three-phase water Raman measurement in the daytime. Furthermore, the obtained overlapping values are applied to accurate retrieval theory for three-phase water profiles. The simulated profiles of atmospheric water vapor, liquid water and ice water indicate that the water vapor, liquid water and solid water content can be increased synchronously in the cloud layer, and their content, distribution characteristics and the corresponding error are also discussed. The above results validate the feasibility of highspectral spectroscopic technique for detecting the synchronous atmospheric three-phase water, and will provide technical and theoretical support for synchronous retrieval of three-phase water by Raman lidar.