The distribution data of fine particulate matter (PM2.5) and ozone (O3) concentrations at high spatiotemporal resolutions play a pivotal role in monitoring and collectively combating atmospheric PM2.5 and O3 pollution, thereby enhancing the quality of human habitation environments. This study integrates reanalysis, satellite observations, and ground-level monitoring data to develop a GLMM-DF (generalized linear mixed effects model-deep forest) hybrid model for joint estimation of PM2.5 and O3 concentrations in the Beijing-Tianjin-Hebei region. The model effectively considers the spatiotemporal heterogeneity inherent in PM2.5 and O3 and their nonlinear relationships with the estimating variables. The cross-validation results indicate that the GLMM-DF model exhibits high estimation accuracy. For PM2.5 estimation, the coefficient of determination (R2), mean absolute error (MAE), and root mean square Error (RMSE) are 0.949, 5.81 μg m−3, and 9.42 μg m−3, respectively. Regarding O3 estimation, the R2, MAE, and RMSE are 0.966, 7.77 μg m−3, and 11.82 μg m−3, respectively. Furthermore, the model demonstrates favorable spatiotemporal performance, with the PM2.5-O3 interaction contributing to enhancing model performance. Utilizing this model for mapping, we analyzed the spatiotemporal trends of PM2.5 and O3 in the Beijing-Tianjin-Hebei region, uncovering several intriguing phenomena.