The computational effort and time required to design the best shapes that can sustain a high stress and deformation level is usually computationally hectic and time consuming. To improve efficiency, there is therefore needs to minimize the design time without affecting the quality. This research paper is based on the use of Tabu Search (TS) algorithm in parameter optimization in optimization of shapes of rubber bumpers in air springs of heavy-duty vehicles. TS is a metaheuristic optimization method that is ideal in multidimensional problems that are complex enough, and this makes it a potential choice in the optimization of engineering designs. The algorithm enhances exploration as it allows exploration to avoid moves that revisit solutions that have already been explored in order to overcome local optima and find globally optimal solutions. In addition, TS has the ability to effectively search the vast spaces, enabling a thorough search of the objective functional landscape. Nevertheless, it has a high computational cost, which is seen in the multitude of evaluations that must be carried out, which is a serious problem. To solve this the current research incorporates a surrogate modeling technique alongside TS that gives a learned predictive model to estimate objective measures. The hybrid strategy minimizes the cost of computation and maintains its accuracy in optimization. The ease of the surrogate model in combination with the TS framework allows the efficient exploration of the design space that produces optimal bumper geometries that are much more computationally efficient and produce better design quality.