Abstarct: Beam-column joints play an important role in transferring bending moments and shear forces between adjacent beams and columns in a bending-resistant frxame. The beam-column joint behavior has been widely investigated by different researchers who consider the parameters effect such as beam and column geometry, concrete and steel grade, beam and column longitudinal reinforcement, joint shear reinforcement and column axial load. The shear strength of reinforced concrete beam-column joints is inevitably affected by various uncertainties that exist in complex shear behaviors. In particular, external beam-column joints (EBCJS) typically transmit higher shear forces but typically have lower shear strengths compared to internal joints. Consequently, beam-column joints (BCJS) are more prone to brittle shear failure, which should not be allowed to occur if a ductile behavior of the frxame is required. To overcome this limitation, a probabilistic prediction model for shear strength of reinforced concrete (RC) beam-column connections baxsed on Gaussian Process Regression (GPR) with heterogeneous combination was proposed. 237 experimental sets of internal joints and 273 external sets of beam-column joints were collected from articles and articles. Then, each of fifteen specific points were affected, including column width, column height, concrete compression, width and point height. It was transverse. The beam, the yield stress of the longitudinal bars of the beam, the middle bars of the column and the horizontal tension, the axial load of the column, the total cross-sectional areas of the column bars and the enclosing circles, the number and diameter of the bars at the top and bottom of the beam section are calculated using Gaussian trend regression and coefficient . The importance of each input factor (IIF) was shown using artificial neural network (ANN). Also, multi-linear, linear, Full Quadratic, interaction, and nonlinear regression models are used to predict shear strength and the greatest impact of input factors on output for existing shear strength models. The prediction performance of each regression model was evaluated using evaluation criteria. The results of the regression models show that the height (hc), the diameter of the upper longitudinal bars of the beam (dtb), the width of the column (bc), the cross-sectional height of the beam (hb), the diameter of the lower bars of the beam point (dbb), the compressive strength of concrete. (fc), the number of rebars at the bottom of the beam section (nbb), the number of rebars above the beam surface (ntb), the yield stress of the rebars along the beam section (b), the horizontal stiffness of the connection (Ash) and the axial load of the column (P) are controlling parameters which have an effect in increasing shear strength.