Quasilinearization is extended to the numerical solution of multi-point boundary-value problems for ordinary differential equations. It is described how such problems arise from state-constrained optimal control problems. A simple numerical example is given to illustrate the method.