Recent research has shown a close temporal relationship between the nocturnal decrease in rectal core temperature and the initiation of sleep. However, there is not yet a clear temporal relationship between changes in peripheral and core temperatures and nocturnal sleep onset. We recorded body temperatures in 14 adult males (age±SEM=22.1±0.6 y), who attended the sleep laboratory for an adaptation night and two counterbalanced experimental sessions. Subjects self‐selected lights‐out on one experimental night (the Habitual Sleep condition). To determine the relationship between body temperature changes and sleep onset, lights out was delayed until after 01.00 hours on the other experimental night (Delayed Sleep condition). Individual datasets in both conditions were expressed relative to the time of sleep onset in the Habitual Sleep condition only, so that they were aligned at identical clock times. Saliva samples confirmed that mean dim light melatonin onset (DLMO) occurred at 00.10±00.16 hours in the Delayed Sleep condition, which was after habitual sleep onset at 23.44±00.08 hours. Rectal core temperature (Tc) decreased significantly over time only in the Habitual Sleep condition ( P < 0.01). For the 20 min before habitual sleep onset, Delayed Sleep Tc was on average 0.1°C higher than Tc in the Habitual Sleep condition ( P < 0.01). The greater decline in Habitual Sleep Tc was associated with significantly increased peripheral hand and foot skin temperatures before sleep (both P < 0.05). Subjective sleepiness measures were higher in the Habitual Sleep onset condition from 150 min prior until sleep onset ( P < 0.01). From these results it is reasonable to infer that a sequence of thermoregulatory and sleep propensity changes occur before, but are associated with habitual sleep onset, as the changes are significantly attenuated if sleep is delayed.