Abstract: For practical autonomous vehicles, the generated trajectories should ensure the feasibility imposed by kinematic, dynamic and actuation. To generate a locally feasible trajectory from the initial state to the target state, a trajectory generation algorithm based on quartic Bzier curve is proposed. Firstly, the original problem is decomposed into shaping the trajectory and executing the shape. To satisfy the kinematic constraints, initial state and target state constraints and continuous curvature constraint, a quartic Bzier curve defined by 3 parameters is adopted to shape the trajectory. To further ensure the feasibility of steering, optimization is utilized to resolve a set of parameters to generate a trajectory that has a minimum curvature change. For velocity generation, an interval of executing time is firstly generated to ensure a continuous velocity, continuous acceleration, bounded acceleration and side-slip avoidance at the target state. Then, the executing time that could avoid side-slip at every point is resolved by adjusting the time. Finally, the executing velocity of the trajectory at each point is generated based on the executing time. To verify the algorithm, trajectory generation based on real environment data and simulations on search space generation, driving on road and path tracking are conducted.