The present study aimed to: (1) investigate sleep architecture in response to an overload training and taper periods among endurance runners; and (2) assess the sleep benefits of a high-heat-capacity mattress topper. Twenty-one trained male endurance runners performed a 2-week usual training regimen (baseline) followed by 2-week overload and taper periods. From overload to the end of the taper period, they were assigned into two groups based on the mattress topper used: high-heat-capacity mattress topper (n = 11) or low-heat-capacity mattress topper (n = 10). Training load was assessed daily using the session rating of perceived exertion. Following each period, sleep was monitored by polysomnography, and nocturnal core body temperature was recorded throughout the night. Irrespective of the group, awakening episodes > 5 min decreased following overload compared with baseline (-0.48, p = 0.05). Independently of mattress topper, each 100 A.U. increase in 7-day training load prior to polysomnographic recording was associated with higher slow-wave sleep proportion (β = +0.13%; p = 0.05), lower sleep-onset latency (β = -0.49 min; p = 0.05), and a reduction in the probability of transition from N1 sleep stage to wakefulness (β = -0.12%; p = 0.05). Sleeping on a high-heat-capacity mattress topper did not affect any sleep variable compared with a low-heat-capacity mattress topper. Increased training loads promote slow-wave sleep and sleep propensity, highlighting the adaptative nature of sleep to diurnal activity and the role of sleep in physiological recovery. Further studies are required on the potential benefits of high-heat-capacity mattress toppers on sleep architecture among athletes.