While complex embedded real-time systems (ERTS) such as robots in operation, there is a possibility that unstructured and unrelated data may be gathered over a period of time through sensors and may result in unexpected behaviors or catastrophes. Without proper modeling of non-deterministic behaviors, implementing highly expected results to handle complex situations is expensive to the designers and may result in numerous programming challenges. For analyzing such situations, a stable and general modeling framework to support the designers for rapid analysis of the system behavior is needed. This paper proposes a generic behavioral modeling framework for embedded real-time systems in uncertain environments based on the few empirical studies. The key contribution of the paper is to develop a framework which can be applied to many ERTS applications, where the system behaviors can be predicted exactly during system in operation. Moreover, the architecture gives overall flexibility to apply all possible behaviors in different situations dynamically. The behaviors are generated by applying various facts and rules which are mapped to tasks. As the limited number of tasks may generate unlimited number of rules and thus unlimited number of behaviors, the modeling architecture provides a best possible way to optimize the necessary behaviors and completely discard the less useful behaviors.