Abstract: A hybrid Active Noise Reduction (ANR) headset system with both fixed-coefficient and adaptive filters is investigated. Besides the broadband noise reduction, the main purpose of the fixed-coefficient filter is to reduce the uncertainty of the real secondary path. Both of the fixed-coefficient filter and the filter to model the real secondary path are optimized with the H₂/H_∞ method. Computer simulations and experiments have been conducted and the results show that the fixed-coefficient filter stabilizes the updating process of the adaptive algorithm and the proposed ANR headset system works well in both the broadband and periodic noise environments.