ABSTRACT In conventional masonry buildings, masonry walls are key structural load-bearing elements. Likewise, masonry infill walls strengthen framed constructions against lateral stress. The material characteristics of brick units and mortar determine the compressive strength of structural masonry walls. In this study, advanced machine learning (ML) techniques were utilized to estimate the compressive strength of structural masonry walls based on the material properties of brick units and mortar. The Young’s modulus of brick units (Eu), compressive strength of brick units (Fcu), Young’s modulus of mortar (Em), and compressive strength of mortar (Fcm) were used as input parameters to model the compressive strength (Fc) of the structural masonry wall. Gradient Tree Boosting (GTB), Elman Neural Network (ENN), and Multivariate Adaptive Regression Splines (MARS) were developed using four diverse input and output (I/O) combinations to explore the effect of each input parameter on output estimation. The data used for modeling were obtained from prior studies published in the literature. The model’s performance was evaluated based on different statistical (error and efficiency) indices. For the third and fourth (I/O) combinations, the MARS model significantly outperformed other models. However, for the first I/O combination, the GTB model performed well. The study also revealed that the compressive strength of a structural masonry wall is more likely to depend on the strength and quality of the brick units than on the strength of the mortar.