The complex shear behavior ofcircular concrete-filled tube (CCFT) members has been a challenge for an adequate design equation. Collapses due to shear failure are primarily seen in shear links, pile foundations, and coupling beams in composite shear walls. The current design provisions are based on limited experimental data, leading to very conservative expressions of shear strength. The recent advances in Artificial Intelligence (AI) technologies provided an opportunity to establish design models directly from the data with no need to postulate a mathematical expression. This study utilized three AI techniques alongside 141 experimental test results from the literature to overcome the complex behavior of the CCFT members by proposing reliable design equations/models. Namely, Gaussian Processing Regression (GPR), Gene Expression Programming (GEP) and Nonlinear Regression (NR) analysis. The predictor variables include axial loading, materials properties, section slenderness ratio and shear span ratio. This paper sheds light on the current data-based techniques in solving complex structural problems by addressing the noted AI methods and their application in predicting the shear capacity of CCFT members. It is concluded that the data-driven proposed model demonstrates remarkable accuracy in predicting shear capacity compared to the current design equations and can be used for routine design practice. The statistical validation resultsshow that among the proposed methods, GPR showed the highest efficiency in predicting the shear capacity of CCFT with an average error of 0.5%, whereas for GEP and NR, average errors are 1.26% and 1.09%, respectively.