The most pressing issue in soil hyperspectral analysis is technical failure due to soil heterogeneity

Hyperspectral technology is an efficient and practical approach for measuring soil properties. However, its predictive accuracy is often limited by spectral variability resulting from soil heterogeneity. Such variability poses challenges for traditional global spectral modeling methods in capturing the highly complex mapping relationships between spectral features and soil attributes. To address this issue, this study introduces a novel spectral similarity-based partitioned modeling strategy and compares it with a soil type-based grouping approach. The proposed method clusters soil samples according to their spectral characteristics to minimize within-group variance. Subsequently, individual inversion models are developed for each cluster, enabling more precise learning of stable relationships between specific spectral patterns and soil properties, thereby improving both model accuracy and interpretability. To validate this methodology, the study systematically evaluates the performance of five traditional clustering algorithms (TCA) and five deep clustering algorithms (DCA) using a national soil spectral library from China, supplemented with local data from Xinjiang, targeting soil organic matter (SOM). Results indicate that both spectral similarity-based and soil type-based grouping strategies significantly surpass conventional global modeling methods. Among them, the spectral similarity-based partitioning approach yielded the most notable improvements: it increased R² values by 14.08% to 30.77% and reduced RMSE by 44.18% to 66.98% compared to traditional modeling. When compared to the soil type-based grouping strategy, it further enhanced R² by 9.46% to 21.43%, while reducing RMSE by 27.23% to 58.20%. These outcomes suggest that conventional soil taxonomic information alone is insufficient to capture critical spectral variations. In the spectral grouping stage, DCA were found to identify and aggregate soil samples with high spectral similarity more effectively than TCA. In particular, the inversion model constructed using Automatic Encoding Clustering(AEC) delivered the best performance. Additionally, the study revealed considerable spectral diversity even among soils of the same type, and no clear correlation was observed between spatial proximity and spectral similarity among samples. These findings confirm that spectral similarity-based grouping and modeling substantially improve the predictive accuracy of soil property estimation, offering a more refined and effective modeling pathway for hyperspectral soil detection.