We study the channel estimation and power allocation method for the Cooperative relaying sys-tem using Non-orthogonal multiple access (CRS-NOMA). Our object is maximizing the Average effective signal-to-interference-and-noise ratio (AESINR) of the weak user while guaranteeing bounded AESINR for the strong user. We transform the nonconvex optimization problem into a Difference of convex (DC) programming and propose a Sequential parametric convex approximation (SPCA) based iterative algorithm to obtain a local optimum. For comparison, we consider a similar optimization problem for a Non-cooperative NOMA (NC-NOMA) system. Simulation results show that the AESINR performance of the CRS-NOMA outperforms signicantly that of the NC-NOMA system.