Organized intelligent communications can effectively improve the surveillance quality for wireless sensor networks. In this paper, we propose a decentralized Belief-Desire-Intention (BDI) oriented system that adopts the BDI mental architecture for wireless sensor networks. It tunes the sensor operational parameters in order to reduce the power consumption upon the satisfactory of the situational awareness quality. In the BDI intelligent architecture, the agent behavior is composed of beliefs, desires, intentions, and actions. Various methods are used in the agents to optimize the initial sensing coverage, network connectivity, senor sleep schemes, and packet transmissions. Simulation results for a special situational awareness application demonstrate the effectiveness of the proposed system.