Guiding navigation is an important application in wireless sensor networks to make moving objects leave dangerous areas safely and quickly. However, guiding policy without considering congestion problem will postpone the escape time. In this paper, we propose a distributed flow-based guiding protocol for indoor environments to evacuate mobile objects from dangerous area to exit. Our goal is to construct less congested paths to minimize the escape time and the congestion time. Each sensor in the network holds an artificial potential value determined by the moving speed of object on a path monitored by it, the distances to exits, and the capacity of exits that have the maximum potential values, and objects are directed to paths with higher potential values. Nevertheless, our approach also solves the local maximum problem that objects are trapped at a non-exit location holding a local maximum potential value. In addition, we adopt a traffic flow model for transforming object density of a road into object's velocity to reduce the computational cost and the communication overhead. Simulation results show our proposed protocol can efficiently decrease the escape time and congestion time.