Global energy consumption in wireless sensor networks restricts the application of the entire networks, including the impact of limited energy capacity of a single node to the system fundamentally. This paper presents a systemic modeling approach for wireless sensor network based on radial basis function neural networks and status-sphere expression. In consideration about the topology and hierarchical structure of WSN, it introduces real-time adjusting of radial basis function neural networks, and establishes matrix model for systematic energy consumption adaptively. Results prove that this model performs effective global optimization by adjusting parameters according to real application circumstances.