Abstract. Human Resting Energy Expenditure (REE) exhibits a robust circadian oscillation, yet current thermal comfort standards and building simulations typically rely on static metabolic assumptions. This preliminary work aims to integrate chronobiological principles into building engineering by developing a dynamic REE model based on accessible biometric parameters. We analyzed continuous 24-hour energy expenditure data from six adults in a metabolic chamber under unstructured, "free-living" conditions. By employing a signal processing pipeline combining Asymmetric Least Squares smoothing with Linear Mixed Models, we isolated the endogenous circadian rhythm from behavioral noise, such as physical activity and the thermic effect of food. Among twelve configurations, incorporating Sex and Body Surface Area was identified as the best-performing model, maximizing statistical fit (Log-likelihood: 88.5) while maintaining a low predictive error (RMSE: 0.109 kcal/min). Results confirm a consistent metabolic peak in the biological afternoon between 13:03 and 17:26. Body Surface Area was the strongest predictor of the metabolic baseline, significantly modulating both the baseline and the amplitude of the circadian swing. While this model targets REE during minimal activity and does not directly generalize to high-intensity tasks, it refines the physiological baseline (0.75 ± 0.08 met ) upon which all other energy costs are superimposed. Accordingly, this time-dependent model provides a method to predict occupant heat generation based on clock time and body composition, supporting more accurate cooling load predictions and human-centric thermal comfort controls.