Overhead gantry crane systems have been integral part of many heavy industrial applications needing material handling and transportation. This can easily be seen in emerging countries growing rapidly. Single force control input applied to crane moves it to a desired position but also triggers undesired swing on the payload which makes it an interesting control problem for this underactuated system. Given an application and a work environment constraint the response time to the position error vs. the maximum allowable amplitude of the payload swing may vary. An intelligent control methodology has been given such an overhead crane system with an optimization parameter compromising between response rate and swing amplitude. The proposed controller does not rely on the knowledge of the crane dynamics making the solution applicable changing operation conditions, so it is a universal controller. Robust adaptive properties are established via a neural network based multi-loop controller architecture. Customization to fit two specific application requirements has been illustrated in the simulations. Stability proofs are given in the sense of Lyapunov. Simulation examples are shown to illustrate the effectiveness of the proposed controller that does not rely on prior knowledge of system dynamics.