In this study, a two-stage iterative optimization algorithm is proposed for articulated automated guided vehicle (also known as tractor-trailer vehicle) trajectory planning in complex static environments. The first stage employs the Generalized Voronoi Graph (GVG) method for map initialization, followed by an improved A* algorithm to generate the initial guess and construct safe tunnels based on this initial guess. The second stage compresses the solution space of the optimization problem to within the tunnels, and an iterative optimization framework is introduced to effectively avoid local optima, rapidly converging to the global optimum. Moreover, this iterative optimization framework is applicable for the optimization of any criterion that can be explicitly definable via a polynomial expression. Compared to traditional algorithms, our algorithm achieves an approximate 45% enhancement in optimization performance.