Abstract: Interference alignment (IA) effectively improves the capacity of interference channel by overlapping the interference at the receivers.At present,most of the distributed interference alignments algorithm design precoder and receiver filter at both transmitters and receivers which causes overhead. In this paper, an improved one-side interference alignment precoding algorithm, which is based on the minimization of the chord distance of the interference subspace.we build the precoding matrix on a Stiefel manifold and make the constraint optimization problem be transformed into unconstraint optimization problem. The simulation results show that the precoding matrix build on the Stiefel manifold compared with the conventional optimization algorithm have the faster convergence rate and higher sum rate.