This paper presents a technique for an intelligent robot to avoid moving obstacles and to reach a goal position in uncertain circumstances. A robot in non-stationary environments is required to infer changing situations. Behavior-based approach is robust to manage interactions within an environment, but in spite of its strengths of fast responses, it has not been much applied to more complex problems for planning or reasoning. This paper proposes a hybrid control architecture to infer dynamic situations. The lower level is to generate reflexive and autonomous behaviors with behavior network, and the higher level is for a mobile robot to infer dynamic situations with Bayesian network. Experimental results with various simulations and a real robot have shown that the robot reaches the goal points while avoiding stationary or moving obstacles with the proposed architecture.