A method for selecting parameters in HgCdTe crystals has been proposed， utilizing Principal Component Analysis （PCA） and clustering methods， with the establishment of a data model for screening the parameters of HgCdTe crystals. Within the model， the initial crystal data undergoes a cleaning and analysis process. PCA is employed for dimensionality reduction， and the Density-Based Spatial Clustering of Applications with Noise （DBSCAN） algorithm is used to identify the densest regions within the crystal data. Furthermore， the high-performance chip data， obtained after post-processing， is utilized to fit boundary ellipses for high-quality HgCdTe crystal parameters. These ellipses act as criteria for identifying high-quality crystals. The model is capable of generating crystal ratings based on input electrical and optical parameters with a coverage rate exceeding 90%.