A　neural　network　(NN)　based　control　method　is　proposed　for　ship　course-keeping　control　in　the　presence　of　modeling　errors.　The　controller　is　constructed　by　combining　both　the　dynamic　surface　control(DSC)　technique　and　minimum　learning　parameter(MLP)　technique　based　on　Lyapunov　stability　theory,　and　the　problem　of　explosion　of　complexity　is　avoided,　so　that　the　computational　burden　of　the　algorithm　can　be　reduced　drastically,　which　is　convenient　for　practice　and　applications.　The　proposed　NN　based　controller　guarantees　that　all　the　close-loop　signals　are　uniform　ultimate　bounded　(UUB)　and　that　the　tracking　errors　converge　to　a　small　neighborhood　of　the　desired　trajectory.　Simulations　illustrate　the　effectiveness　of　the　proposed　algorithm.