Animal body temperature rises throughout the day and peaks in the evening, a pattern conserved across diurnal endotherms and ectotherms. However, the mechanisms driving the robust body temperature rhythms (BTR) remain largely unclear. Here, we developed a machine learning-based platform, temperature homeostasis evaluation of rhythmicity in model Drosophila (THERM-D), enabling continuous, high-throughput analyses of BTR. Using THERM-D, we identified robust BTR patterns reflecting flies' morning and evening behaviors and revealed the function of CRYPTOCHROME (CRY)-negative clock neurons. About half of all clock neurons lack CRY, yet their function was unclear. Newly developed Gal4 drivers targeting CRY-negative neurons demonstrated that these neurons control the morning temperature rise without affecting evening BTR. The data suggests that separate clock circuits regulate morning and evening BTR. Thus, THERM-D elucidated the role of CRY-negative clock neurons, which are specialized for BTR regulation and distinct from the circuits controlling sleep-wake cycles.