This study proposes a novel outlier detection method for beta regression models, known as the BP-Pearson residual method. This method combines Tukey’s boxplot with Pearson residuals to enhance outlier identification. Its effectiveness is assessed through simulation studies on both uncontaminated and contaminated datasets, focusing on three performance metrics: the probability of accurately detecting all outliers, the probability of outliers being falsely classified as inliers (masking effect), and the probability of inliers being identified as outliers (swamping effect). Real data are also used to demonstrate the BP-Pearson residual method’s performance. Results from both simulations and real data applications indicate that the BP-Pearson residual method is effective for detecting outliers in beta regression models, showing robust performance across various data scenarios.