Path planning is an important branch in the field of robotics, and the use of Signal Temporal Logic (STL) to solve planning problems is currently a very hot research topic. STL, as a logic language for describing task specifications and executes them, can express various spatiotemporal constraints. Its semantic robustness is used to analyze systems over a certain continuous time, by proposing corresponding robust functions to quantify the degree to which a system satisfies or violates the STL specifications. In the environment with dynamic and static obstacles, based on the Real-Time Rapidly-Exploration Random Tree (RT-RRT*) planning algorithm, a robust function based on STL and a smoothing function are proposed to maximize the system’s robustness and smoothness, thereby finding the optimal solution. Simulation results indicate that this method can compute and compare multiple cost values to select the most robust planning path, satisfying the STL specifications.