The intention recognition problem is a difficult problem which consists in determining the intentions and the goals of an agent. Solving this problem is useful when they are several agents which are interacting with each other and when they do not know each other. The effectiveness of an agents’ work could be improved in this case. We present a method to infer the possible goals of an agent by observing him in a series of successful attempts to reach them. We model this problem as a case of concept learning and propose an algorithm to produce concise hypotheses. However, this first proposal does not take into account the sequential nature of our observations and we discuss how we can infer better hypotheses when we can make some assumption about the behavior of the agents and use background knowledge on the dynamics of the environment. We then provide a simple way to enrich our data by assuming the agent can compute the effects of his actions in the next step and study the properties of our proposal in two different settings. We show that our algorithm will always provide a possible goal if such a goal exists (meaning that there is indeed some set of states in which the agent always succeeds and stops in our observations).