The predicted exponential increase in data traffic for future 5G networks demands for increased wireless spectrum capacity. The unlicensed spectrum bands are overburdened, whereas the licensed spectrum bands are underutilized. Cognitive radio systems (CRSs) can solve the problem of spectrum scarcity by allowing the reuse of the underutilized licensed spectrum bands. However, efficient resource allocation schemes in CRS are inevitable before we can reap the benefits of CRS. In this article, we have formulated an optimization problem for multiuser cooperative CRS, which considers relay selection and power allocation to maximize the sum capacity of the system. The optimization problem is a mixed‐integer nonlinear program, and deriving its optimal solution is extremely difficult. Therefore, we propose an iterative joint multiple relay selection and power allocation (IJRSPA) algorithm for multiuser CRSs. The proposed IJRSPA has low computational complexity, and simulation results verify its effectiveness.