In industrial control systems, more than 90% of control loops belong to the PI or PID type.　We compared several well-known PID tuning laws in this thesis. There are simple tuning formula to design PID controllers for 4th-order with the dead-time plant. However both the performance and robustness of those well-known PID tuning formula for the long dead-time process, delay time to time constant is large than 1, are not acceptable.　Moreover, the robustness specification was not be considered in some PID tuning laws. Therefore, we used the SPSA random search methods to solve these problems. The first step of this method is to choose the solution space, and then to search the optimal solution by specially designated stochastic rules with an optimal objective function constrained with limited conditions. An auto-tuning PID controller for position control of a voice coil motor is developed in this thesis. The on-line parameter tuning policy is based on the SPSA technique. Three PID parameters can tuned consecutively until the optimal values are attained. The proposed approach has been confirmed in computer simulations and experiments. The optimal search method was implemented in the position control for a voice coil motor.