A trajectory generation method for car-like mobile robot based on cubic spirals and line segments is presented. The generated path is made up of portion of cubic spiral segments with zero curvature ends and upper bounded curvature and straight line segments. A numerically efficient process, which is resorted to minimization over the sum of length of each path segment of generated path via linear programming, is presented to generate a Cartesian shortest path linking start and destination configurations of car-like mobile robots through an intermediate configuration. The intermediate configuration is not necessarily selected from the symmetric means circle. The merits of our path generation method based on cubic spirals are: (i) The implementation is straightforward so that the generation of feasible paths with bounded curvature is efficient for real-time applications. (ii) Applicable to mobile robots with forward and backward driving abilities and only forward driving ability; (iii) Flexible to incorporate other constraints.