Models obtained in our previous studies were adapted from a linear model. In these models, the operative temperature has a positive coefficient with respect to the average heart rate during sleep (SHR) in summer and a negative coefficient in winter. Therefore, there are limitations to using linear models to examine annual variations. This study attempted nonlinear modeling and conducted an analysis. The participants were nine healthy males in their twenties, with an average body mass index, living in a temperate region. The objective variable was SHR. The explanatory variables were the body movement rate during sleep (BM), room temperature, radiant temperature, bed microclimate temperature, and carbon dioxide concentration. In addition, the interaction effects between the bed microclimate temperature and CO2 concentration were considered because they showed significant correlations in the linear model. Results showed that the BM and SHR exhibited a positive linear correlation throughout the year. Radiant temperature showed a higher importance than room temperature in winter, as evaluated by permutation importance. SHR was lowest when room and radiant temperatures were approximately 25 [℃] to 26 [℃]; under higher or lower temperatures, the SHR increased. Additionally, when temperatures fell below 10 [℃], the trend shifted to a positive correlation. Bed microclimate temperature and SHR showed a positive correlation when exceeding 30 [℃]. The CO2 concentration and SHR exhibited a positive correlation below 1000 [ppm] during autumn and winter. Overall, a greater understanding of how environmental conditions affect SHR could enable the design of environments that promote a lower heart rate during sleep.