This work deals with the localization and mapping problem of a car-like mobile robot in an unknown urban environment. The robot is equipped with a Sick LMS511 Pro laser sensor, the laser data are exploited to accomplish the task of localization and mapping. In such a situation, the two problems can not be dissociated, therefore, we propose to adopt a Simultaneous Localization And Mapping (SLAM) approach. The developed solution is called NDT-PSO. It is based on the Normal Distribution Transformation (NDT) method and the Particulate Swarm Optimization (PSO) method. NDT is a SLAM method whose principle is to determine the geometric transformation between two successive laser scans using alignment techniques. Its representation of the environment is based on the modeling of the all 2D points reconstructed from a laser scan by a collection of local normal distributions. The PSO method is used in the transformation parameters optimization step , this parameters are used to determine the poses (positions and orientations) of the robot. The proposed algorithms are implemented on the Robot Operating System ROS using Python language and tested on the RobuCar mobile robot as part of an intelligent urban transport project.