With the deepening understanding of complex systems, traditional network models based on pairwise interactions have exhibited limitations in describing multi-body processes, such as the spread of epidemics within populations.Consequently, higher-order network representations based on simplicial complexes have become increasingly necessary. However, most existing epidemic models on higher-order networks assume constant transmission probabilities, failing to capture the dynamic changes induced by individual behavioral adaptations or environmental shifts during the spreading process. To address this issue, this paper proposes an SIR epidemic model operating on second-order simplicial complexes. The core innovation lies in a Bayesian dynamic updating mechanism based on mean-field approximation: by treating local transmission events within the network as approximately independent Bernoulli trials and leveraging the Beta-Binomial conjugate property, the model dynamically updates the posterior estimates of first-order (edge) and second-order (triangle) transmission probabilities using real-time infection data. Extensive independent Monte Carlo simulations conducted on Erdős-Rényi (ER) random graphs, alongside rigorous comparisons with fixed-parameter models possessing equivalent average transmission capabilities, demonstrate that: (1) the Bayesian dynamic model sensitively captures the time-varying nature of parameters during epidemic evolution, while the second-order interaction structure significantly accelerates the outbreak and elevates the infection peak; (2) the derived analytical solution of the time-varying basic reproduction number, R₀(t), is highly consistent with the effective reproduction number obtained via simulations, thereby validating the accuracy of the theoretical derivation. This study reveals the coupling mechanism between higher-order interactions and adaptive transmission probabilities, highlighting the importance of incorporating dynamic parameter updates into models to enhance predictive realism and formulate precise prevention and control strategies.