Boresight and jitter are two fundamental pointing errors for a laser pointing system. With the theory of maximum-likelihood estimation, a laser pointing system model is setup and consummated based on a Gaussian far-field irradiance profile and a Gaussian beam jitter model. The estimates are gained by use of return photon counts reflected from an intended target. Then, a Monte Carlo simulation is programmed and a laboratory experiment is performed. Both the simulations and experimental results demonstrate that the performance of the maximum-likelihood estimator is excellent and improves with the increasing number of shots. With this method, the boresight and jitter can be obtained simultaneously. What′s more, the further study finds that the experimental results agree well with the simulation results. Based on these excellent performances of the estimator, a closed-loop laser pointing experiment is exhibited in laboratory.