Abstract: A new approach is presented to studying QoS-aware service combination. An interference model is proposed to quantitatively characterize and measure how the interference affects the service provided for users. The interference intensity is defined as the substituted number of nodes per unit time when recovering a service path. To reduce the service routing computation, an efficient routing algorithm is designed, with three optional strategies for simplified routings provided, i.e., lazy adjustment, selecting reliable nodes and local path recovery. Then, based on the three strategies and the interference model proposed, an optimized minimum interference service routing algorithm is given. The results showed that both the algorithm and model proposed can reduce the interference on end users and improve the system robustness and stability when finding service paths.