To accurately estimate the shear strength of RC (reinforced concrete) shear walls with boundary elements and to reasonably describe its probabilistic characteristics, a probabilistic model for shear strength of RC shear walls was proposed based on an improved strut-and-tie model and Bayesian theory. Firstly, the contribution of longitudinal reinforcement in the boundary elements was incorporated into the softened strut-and-tie model, leading to the development of a probabilistic model that accounts for the boundary element effects. Then, using a dataset of 392 experimental specimens, the statistical characteristics of the model parameters (including empirical distribution types, means, standard deviations, and correlation coefficients) were systematically determined through Bayesian theory and MCMC (Markov Chain Monte Carlo) method. Finally, the proposed model was validated by comparing its predictions with experimental results and several traditional models. The analysis demonstrates that the proposed model not only effectively captures the probabilistic features of shear strength but also enables the calibration of conventional models using confidence intervals and probability density functions. Meanwhile, the proposed mean-value model shows strong agreement with the actual shear strength, with a mean and standard deviation of the predicted-to-measured ratio of 0.95 and 0.31, respectively.