Pervasive devices interacting in open and dynamic spaces with each others require a mechanism that allows them acting autonomously in a secure way and protecting their resources. Trust is fundamental to establish communication with other users, because the identity is often uncertain and on one's own does not provide trust information, for instance, could an unknown user be trustworthy? Nowadays, these devices have a so limited security support. So, we propose a simple trust management model to enhance such support, allowing them interact in ad hoc networks and peer-to-peer applications in a secure way. In this paper, our main contribution is a mathematical and a probabilistic model, as well as demonstrating the model feasibility, since it has been assessed through the prototype implementation, which has been tested in a Pocket PC.