According to the U.S. National Highway Traffic Safety Administration, single vehicle road departures result in many serious accidents each year. An intelligent lane markers recognition and localization system can assist vehicles stay in proper location of a lane that will reduce possibility of car accidents correspondingly. A great deal of lane recognition algorithms have been developed over the past several decades. However, reliable detection is still an issue because of variable road face conditions. An optimum algorithm, Probabilistic Hough Transform (PHT) is presented in this paper for intelligent Lane markers recognition which use a fixed camera installed on the vehicle to transmit video information. The result of experiment proved that under inconsistent illumination and a diversity of road conditions, the accuracy and efficiency of developed system have been improved greatly.