A four-dimensional variational (4DVAR) data assimilation system was developed for a regional chemical transport model (CTM). In this study, we applied it to inverse modeling of CO emissions and mineral dust emission flux over East Asia, and demonstrated the feasibility of our assimilation system. In CO inverse modeling, three ground-based observations were used for estimating CO emission over East Asia. Assimilated results showed better agreement with observations; the RMS differences were reduced by 16-27%. CO emission over industrialized east central China between Shanghai and Beijing has increased markedly, and the results show that the annual anthropogenic (fossil and biofuel combustion) CO emission over China are 147 Tg. In dust inverse modeling, NIES LIDAR observations were used. The assimilated results better reflects the presence of the elevated dust layer and improved the under-prediction of dust concentrations. We obtained an 18% increase in calculated dust emissions through data assimilations, especially over the Mongolian region, indicating that the observed high-dense dust layer might originate in that region. These data assimilation results indicate that the 4DVAR method is very powerful for unification of observation and numerical modeling by CTM.