Complex relationships exist between the indoor environment of residences and their occupants' health. However, it is difficult to describe such relations by the use of normally employed regression-based models. The aim of this study is to provide an approach to examine these relationships, making use of structural equation modeling to describe the interaction between indoor environment and health based on data from a questionnaire survey in China. In the model, the indoor environment was divided to the living room, bedroom, kitchen and bathroom. The model also took into account non-environmental factors, such as socioeconomic status and lifestyle. Multigroup analysis was carried out to test for variations arising from different climate zones and lengths of residence. The results showed that the resulting model had satisfactory levels of reliability and validity, and its multi-group invariance was confirmed. The path coefficients of overall indoor environment, socioeconomic status, and lifestyle on the occupants' health were 0.144, 0.560, and 0.191, respectively. The respective weights of the bathroom, bedroom, kitchen, and living room to the overall indoor environment were 0.445, 0.274, 0.261, and 0.138, respectively. The main indicators that delivered the greatest weights to the bathroom, bedroom, kitchen, and living room were unnatural posture (0.294), thermal environment in winter (0.288), scald (0.307) and slip (0.403), respectively. The resulting model therefore describes aspects of the complex relationships between the residential indoor environment and residents’ health, which is helpful for developing means of improving indoor environment quality.