This paper advances a method to implement the on-machine measurement (OMM) with the touch-trigger probe, also called switching probes. Some of the advantages and disadvantages for touch-trigger probe are discussed. However, the touch-trigger probe errors exist and become one of the major errors for the measurement accuracy. Major factors that influence the probe measurement have been analyzed. The basic technique of probe measurement error modeling with artificial neural network was researched, and also the probe measurement error compensation with 3-layered backpropagation artificial neural network was presented. At last in the experimental system composed of DIXI 50 machining center, Fanuc 16i control system, Blum CNC P82.046 probe and PC, valid the correlated techniques. In addition, the connection and communication between the machining center equipped with probe system and the computer have been introduced. The experiment indicated that, using the touch-trigger probe makes on-machine measurement more automatic and efficient. And by using the back-propagation neural network for error compensation make on-machine measurement more precise.