This paper aims to improve trusts in multi-agent systems by proposing the Triple-R model- a computational trust model based on reliability, reputation, and risk. Existing trust models in multi-agent systems were analyzed before integrating the three components into a single Triple-R model to select the most trustworthy provider using TOPSIS. We provide experimental results in defense of our approach and demonstrate the accuracy of the presented model in evaluating the trustworthiness of agents. Results indicate the performance of Triple-R to be significantly better than other existing models, and therefore it is capable of evaluating the trustworthiness of agents more accurately compared to the existing models. Triple-R was also found to be the only model to have improved accuracies with a growing number of providers. The findings indicate that reliability, reputation and risk can be efficiently integrated to enhance trust between agents in a multi-agent environment.