Abstract: Based on overlay-underlay spectrum sharing model, a routing algorithm for hybrid cognitive radio networks is proposed to improve the existing spectrum sharing model and to solve the deficiency in processing interference and delay of other algorithms. Taking shortest path and link state as routing metric and minimum cumulative interference as channel assignment target, the algorithm uses coloring graph as routing analysis model to develop an end-to-end path selection and channel assignment method with power control. Simulation results and comparison with existing routing algorithm show that the proposed algorithm can improve network throughput, reduce packet loss rate and end-to-end delay.