Evidence suggests that there are positive effects of exercise on learning and memory. Moreover, some studies have demonstrated that forced exercise plays the role of a stressor. This study was aimed at investigating the effects of different timing of exercise and exercise withdrawal on memory, and serum and hippocampal corticosterone (CORT) levels. Wistar rats were randomly divided into five groups: control, sham, exercise-rest (exercise withdrawal), rest-exercise (exercised group), and exercise-exercise (continuous exercise). Rats were forced to run on a treadmill for 1 h/day at a speed 20-21-m/min. Memory function was evaluated by the passive avoidance test in different intervals (1, 7 and 21 days) after foot shock. Findings showed that after the exercise withdrawal, short-term and mid-term memories, had significant enhancement compared to the control group, while the long-term memory did not present this result. In addition, the serum and hippocampal CORT levels were at the basal levels after the rest period in the exercise-rest group. In the rest-exercise group, exercise improved mid- and long-term memories, whereas continuous exercise improved all types short-, mid- and long-term memories, particularly the mid-term memory. Twenty-one and forty-two days of exercise significantly decreased the serum and hippocampal CORT levels. It seems that exercise for at least 21 days with no rest could affect biochemical factors in the brain. Also, regular continuous exercise plays an important role in memory function. Hence, the duration and withdraw of exercise are important factors for the neurobiological aspects of the memory responses.