Wireless Sensor and Actuator Networks (WSAN) are becoming more and more relevant and are likely to spread everywhere, due to their numerous applications and benefits. In fact, they are a key element in the vision of the so called Internet of Things. Nevertheless, the lack of central control and the existence of incentives for cheating lead to the necessity of defining mechanisms for identifying well-behaving WSANs. Reputation based solutions have been suggested in the literature in order to cope with these situations, but most of them were focused on finding a specific most trustworthy sensor, capable of provisioning a requested service. Our work, however, is oriented to provide a mobile user with an accurate reputation-based mechanism allowing her to select, amongst a set of available WSANs, the most reputable one, according to the recommendations given by other users. Thus, the mobile user can build her own opinion and consequently decide which is the best WSAN to interact with. This decision will be based in her own criteria, which might be different from other users of the same service. In order to prevent the abuse of the proposed reputation system, we suggest a mechanism avoiding multiple recommendations from the same user in a specific time frame. Additionally, we handle certain privacy issues by hiding the real identity of a user behind a group. The experiments performed show the accuracy, scalability and robustness of our proposal, called WSANRep, within several scenarios, considering the existence of malicious recommenders and malicious service providers.