Hybrid power systems combining fuel cells with fast energy storage devices are good solutions to the fuel cell load-following problem. Hybrid systems may also offer efficiency and reliability advantages. In this paper, we propose a power control scheme for hybrid systems that exploits feed-forward information about the steady-state behavior of incoming load transients. The method uses a modified cluster-weighted modeling (CWM) algorithm to build a load transient recognition model. The model is formulated sequentially and can provide useful feed-forward information in real time. Simulation and experimental results are provided that demonstrate the effectiveness of the transient recognition model and the proposed power control scheme for hybrid fuel cell systems.