The accurate prediction of shallow foundation settlement (Sm) in non-cohesive soils is challenging due to their heterogeneous nature and the numerous parameters involved. To address these difficulties, researchers have sought more precise methods. The utilization of the rock engineering system (RES) method in this paper aims to address the intricacies and uncertainties present in soil and rock properties. This method adeptly models the non-linear and complex relationships among various factors with precision, eliminating the need for intricate coding and resulting in minimal errors. A dataset comprising 189 data points is employed to construct the Sm model using the RES method. Within this dataset, 151 points (80% of the data) are dedicated to model construction, while the remaining 38 points (20% of the data) are reserved for data evaluation. The parameters considered in this study include the width of footing (B), ratio of footing embedment (Df/B), count of SPT test blow (N), geometry of footing (L/B), and pressure of footing (q), all of which affect Sm (mm). Furthermore, regression methods are employed to compare their performance against the RES method. The RES-based method demonstrates superior accuracy and reduced error in comparison to other regression methods, as indicated by three RES statistical indicators (R2 = 0.956, MSE = 0.0045, RMSE = 0.0677). Consequently, the results suggest that the RES method can serve as a valuable and practical tool for engineers specializing in rock engineering.