<p>This article proposes a path planning methodology for autonomous vehicles using a tentacle algorithm and constrained optimization. The tentacle algorithm generates multiple path candidates by varying steering angles. The optimal tentacle path is selected based on the evaluation of a cost function and constraints related to safety and lane-centering behavior. Subsequently, a constrained optimization problem is formulated using the sequential quadratic programming (SQP) method to refine the driving path. The tentacle path serves as a guide for the optimization process. Specifically, lateral safety envelopes are determined by inspecting obstacles in bilateral directions from the tentacle path. This approach efficiently selects a suitable driving corridor among multiple candidates. Moreover, the constrained optimization stage compensates for the limited path-shaping capability of the tentacle algorithm. The performance of the proposed algorithm was evaluated through simulation studies in cluttered environments. The simulation results demonstrate that the proposed integration of the tentacle algorithm and constrained optimization determines a safe and flexible driving path with a low computational burden.</p>