In this paper, we propose a cooperative decoupling method and a cross-layer joint method for multi-layer resource allocation in multi-hop cognitive radio networks. In cooperative decoupling method, the task of path choosing is accomplished independently, and then the game of channel and power allocations is implemented. In cross-layer joint method, the three-layer resource of path, channel and power is allocated simultaneously by process of game. The heuristic principles of network layer, media access control layer and physical layer are employed synthetically in two methods. The degree of receiving interference and the degree of sending interference are adopted to assist path choosing. The Boltzmann exploration based on the width of permitting power is designed to select the channel and power. The means of replacement and elimination of long link or bottleneck link are proposed to further enhance network performance. The sequential game process instead of simultaneous game process is chosen because the former has better convergence property in current scenario, and the concrete process of game is provided. Moreover, the Nash equilibrium of the games and the complexity of the algorithms are analyzed and discussed. Simulation results show that the cooperative decoupling method and the cross-layer joint method have better performances in the number of success flows, the achievable data transmission rate and power consumption than the cooperative link game and the local flow game with simple decoupling.