Conventional wisdom suggests mid-task rest as a potential approach to relieve the time-on-task (TOT) effect while accumulating evidence indicated that acute exercise might also effectively restore mental fatigue. However, few studies have explored the neural mechanism underlying these different break types, and the results were scattered. This study provided one of the first looks at how different types of fatigue-recovery break exerted influence on the cognitive processes by evaluating the corresponding behavioral improvement and neural response (EEG power spectral) in a sustained attention task. Specifically, 19 participants performed three sessions of psychomotor vigilance tasks (PVT), with one session including a continuous 30-min PVT while the other two sessions additionally inserted a 15-min mid-task cycling and rest break, respectively. For behavioral performance, both types of break could restore objective vigilance transiently, while subjective feeling was only maintained after mid-task rest. Moreover, divergent patterns of EEG change were observed during post-break improvement. In detail, relative theta decreased and delta increased immediately after mid-task exercise, while decreased delta was found near the end of the rest-inserted task. Meanwhile, theta and delta could serve as neurological indicators to predict the reaction time change for exercise and rest intervention, respectively. In sum, our findings provided novel evidence to demonstrate divergent neural patterns following the mid-task exercise and rest intervention to counter TOT effects, which might lead to new insights into the nascent field of neuroergonomics for mental fatigue restoration.