Currently, the source location privacy (SLP) becomes a hot research interest in network security of Underwater Acoustic Sensor Networks (UASNs), and existing schemes are mostly proposed for a given scenario. Introducing source location privacy technologies inevitably increase the energy consumption of nodes, while they are widely deployed in available studies, resulting in massive energy wastage. Therefore, an adaptive scheme for protecting source location privacy (APSLP) in UASNs is proposed. The APSLP scheme first analyzes the possible locations of the adversary by trust method. Then, considering the lagging nature of the trust method, which means that the adversary may not stay in locations given by trust method, a hidden Markov-based backtracking method is proposed and location privacy methods are functioned according to the backtracking result. The simulation shows that even though the security level of the APSLP scheme is not the largest, the efficiency is the highest, approximately an increase of 69.1$\%$and 10.3$\%$compared with two comparison algorithms, respectively.