Anticipating the timing of behaviorally relevant events is crucial for organizing movement. The time to initiate actions based on events (i.e., reaction time (RT)) is a useful measure to quantify states of anticipation. Few studies have examined how anticipation affects the timing of limb movements. We addressed this question behaviorally with two macaque monkeys performing delayed wrist movement tasks. The interval between target onset and go signal (i.e., foreperiod) varied randomly from 1 to 2 s. The probability that the go signal was about to occur (i.e., hazard rate) increased as the foreperiod increased. The kinematics of wrist movements was not influenced by foreperiod duration. Analyzing RT data with the LATER model indicated that RT distributions swiveled on reciprobit plots as foreperiods increased, suggesting that changes in RT distributions were due to changes in anticipation. RT was inversely related to hazard rate. To better understand the general implications of anticipatory states, we introduced an additional rectangular foreperiod distribution that ranged from 0.9 to 1.5 s. For that distribution, the hazard rate peaks were higher than those of the 1-2 s distribution. Changes in RT were clearly explained by quantitative differences in hazard rate. The decrease in RT in the 0.9-1.5 s foreperiod distribution was greater than that in the 1-2 s foreperiod. Thus, monkeys learned the temporal structure of foreperiod distributions and anticipated the onset of the go signal, based on hazard rates.