Soil texture is a fundamental parameter that governs processes in agriculture, water resources management, environmental conservation, and land use planning. Recent advances in remote sensing technologies, particularly hyperspectral imaging, have significantly enhanced the ability to characterize and map soil texture with high precision. Hyperspectral sensors capture detailed spectral data that reveals subtle variations in soil properties, often undetected by conventional methods. The present study evaluates the potential of EnMAP hyperspectral imagery, combined with machine learning (ML), to predict and map soil textural classes in semiarid environments. Three scenarios were assessed: (i) using hyperspectral data alone, (ii) incorporating spectral indices, and (iii) integrating terrain parameters. Four machine learning algorithms, Random Forest (RF), XGBoost, Support Vector Machines (SVM), and k-Nearest Neighbors (KNN), were evaluated for soil textural classification. Among them, SVM consistently outperformed the others across nearly all scenarios, achieving an overall accuracy (OA) and precision of 75% and 79.2%, respectively, under Scenario 1 (hyperspectral bands only), which is identified as the recommended configuration for operational soil textural class mapping using EnMAP data. RF and XGBoost also demonstrated strong performance, confirming their robustness for soil textural classification and mapping. In contrast, KNN yielded the lowest classification accuracy, particularly when spectral indices were included, suggesting limited effectiveness in this context. The incorporation of terrain features improved classification accuracy for kNN model, highlighting the value of multi-source data integration. The obtained findings demonstrate that EnMAP hyperspectral imagery, with specific ML, offers a powerful tool for precise soil textural classes mapping, critical for sustainable agriculture, erosion control, and water management. This approach can provide policymakers and farmers with valuable insights for making data-driven land-use decisions that promote soil health and food security.