Animate beings show adaptive behaviors in response to an environment which restricts the individual's movement. If we can model the complex behaviors of animate beings through their analysis, the model can be applied to the design of an artifact which has error tolerance. In this study, we focus interest on fish locomotive behavior in a water area as the so-called the edge region which restrains the fish's behavior. Next, we developed a fish-like behavior algorithm for a mobile robot (e-puck) and compared fish behavior with that of the robot. We adopt chaos, fractal and information content analysis for the water tank experiments and mobile robot experiments. We can find similarities in behavior patterns and in numerical analysis of both. These results offer promising guidelines for discussion of modeling adaptive behaviors of fish and its biomimetical application.