HoneyBadger-Byzantine Fault Tolerant (HB-BFT) is a pioneering practical asynchronous BFT algorithm, distinguished by its lack of reliance on a global clock and its robustness, making it an ideal consensus mechanism for blockchain applications in wireless IoT environments. This paper presents a theoretical model to evaluate the consensus reliability of the HB-BFT algorithm in wireless networks. It begins by analyzing node reliability and failure rates at various consensus phases using conditional probability, then simplifies the expressions to derive a general formula for consensus failure rates. The model further calculates the mean values of reliability and failure rates, revealing a logarithmic linear relationship with node reliability and the number of failure nodes. Corresponding simulations model the impact of node failure on the consensus failure rate. The results indicate that within the tolerable range, the probability of consensus failure increases exponentially with the number of failure nodes.