This paper investigates the optimal entry and exit decisions under a mean-reverting process over a finite horizon. Many theoretical studies on real options assume that an underlying risk follows a geometric Brownian motion over an infinite-time horizon. This assumption is not always practical, especially in discussing realistic investment strategies. In this paper, we examine effects of the mean-reverting process on both entry and exit decisions over a finite horizon. We focus on deriving the optimal boundaries of entry and exit decisions under a mean-reverting process, and compare the effects of the underlying risk process and length of the project horizon on the optimal decisions. Numerical examples in this paper demonstrate that the length of horizon could have a significant impact on the boundaries of the optimal decisions, and hence on project values, particularly under a mean-reverting process.