Unfolding technology can partially alleviate the state explosion problem in Petri net through branching processes. However, an unfolding still contains all states of a system. Some practical problems only need to determine the coverability of a specific state, which is expected to reduce the scale of unfolding net. This study proposes a target-oriented reverse unfolding algorithm for the coverability problem of 1-safe Petri nets, which combines heuristic technology to reduce the scale of unfolding nets, thereby improving the efficiency of coverability determination. Furthermore, the reverse unfolding technology is applied to the formal verification of concurrent programs, and the data race detection problem is converted into the coverability problem of a specific state in 1-safe Petri nets. The experiment compares the efficiency between forward unfolding and reverse unfolding in the coverability problem of Petri net. The results show that when the Petri net has more forward branches than backward branches, reverse unfolding is more efficient than forward unfolding. Finally, the key factors influencing the efficiency of reverse unfolding are analyzed.