AIM: To evaluate the application effectiveness of a Bayesian mixture model based on the principal stratum strategy for estimating the complier average causal effect (CACE) in clinical trials with non-compliance. METHODS: Using a non-inferiority randomized controlled trial investigating a novel drug for primary type 2 diabetes mellitus (non-inferiority margin: -0.4) as a case study, the primary analysis applied a Bayesian mixture model under the monotonicity assumption to estimate CACE of between-group differences in glycated hemoglobin (HbA1c) changes within the compliant stratum, followed by non-inferiority testing. Sensitivity analyses included a Bayesian mixture model relaxing the monotonicity assumption and comparing results with per-protocol set (PPS) analysis. RESULTS: In the primary analysis, the posterior mean of CACE for HbA1c change in the compliant stratum was 0.081%, with a one-sided 97.5% credible interval lower bound of -0.124, exceeding the non-inferiority margin (-0.4%), supporting the non-inferiority efficacy of the novel drug in the compliant stratum (P(H1|Data) = 1). Consistent findings were observed in PPS analyses (estimated effect: 0.136%; one-sided 97.5% credible interval lower bound: -0.069%), further validating methodological robustness. CONCLUSION: In clinical trials with noncompliance as an intercurrent event, the Bayesian mixture model under the principal stratum strategy effectively adjusts for compliance-related bias and yields conservative, robust estimates of causal effects, supporting its value in efficacy evaluation under complex compliance scenarios.