Modular self-reconfigurable (MSR) robots have the ability to change their structure through processing, communication, and movement in a process called self-reconfiguration. Although dynamic self-reconfiguration has shown promising results in different applications, robots still face uncertainties while performing tasks in the real world. For example, they may encounter unexpected obstacles that can divert them from their goals or even cause the system to shut down. This paper introduces a new approach to help robots detect and adapt to obstacles, allowing them to continue achieving their original objectives. This method focuses on the movement and direction of two-dimensional self-reconfigurable robots. We use VisibleSim to test and evaluate this algorithm in different scenarios, from dealing with a single obstacle to navigating a full maze.