Spatial autocorrelation is an important epidemiological feature of hand, foot, and mouth disease (HFMD). However, few studies have included this feature in the regression relationship between HFMD incidence and driving factors to explore its impact on incidence. In this paper, we propose a mixed geographically and temporally weighted autoregressive (MGTWAR) model to explore the impact of spatial autocorrelation and meteorological factors on the incidence of HFMD among children in Inner Mongolia, China, in 2016. In addition, we proposed a residual-based bootstrap test to identify the spatial autocorrelation in the incidence of HFMD and the spatiotemporal heterogeneity in regression relationships. The analysis results indicate that simultaneously modeling the spatiotemporal heterogeneity and spatial dependence of the incidence of HFMD can effectively improve the fitting effect of the model in terms of [Formula: see text]. The MGTWAR model, compared with the GTWAR model, can maintain a simpler model structure while achieving a relatively smaller loss in terms of fitting accuracy, thus having better interpretability. The incidence of HFMD among children in neighboring counties in the Inner Mongolia region shows a noteworthy positive spatial autocorrelation characteristic. Furthermore, this spatial autocorrelation exhibits considerable variation across different regions and months. The impacts of air temperature (AT), air pressure (AP), and average wind speed (AW) on the incidence of HFMD have significant spatiotemporal heterogeneity, and relative humidity (RH) has a global positive influence on HFMD incidence. On the whole, the degree of influence of meteorological factors on the incidence of HFMD is in the order of AT > AP > RH > AW, and the influence of spatial dependence of the incidence of HFMD can not be ignored when exploring the driving factors of HFMD incidence and formulating preventive measures.