With the development of mobile devices and wireless communication technologies, users and devices will face an environment full of changes. It is indispensable for a system to adapt to the environment changes, which we call context-awareness. This paper introduces the adaptive mechanisms for embedded systems from the point of view of the data. The data in a context-aware system are in three different forms in different stage: the sensor data, the context, the situation. This paper introduces the context model which consists of three layers and the adaptive mechanisms in different layer to enable the system adaptive to contexts. Besides, we develop an example to show how the clustering and reasoning technology are used to recognize and derive the higher level context.