In this thesis, the research, design and experimental examination of different control structures on electro pneumatic servo system have been presented. The mathematical models of a pneumatic rodless cylinder controlled by three types of valves: proportional directional control valve, proportional pressure regulators and on/off solenoid valves are given. The dynamic models were built on the basis of theoretical considerations and of experimental data. A linear PID feedback throttle position controller is algebraically optimized according to the damping optimum. The PID controller is tuned to provide critically damped (aperiodic) response of the throttle position control loop. In order to deal with different parameters of the linear process model for the regions below and above the limp-home position, a PID controller gain-scheduling algorithm is applied. The proposed feedback controller is extended with a feed-forward controller placed in the position reference path, in order to obtain faster throttle reference response. In order to analyse and compare the usefulness of control structures of position control system with each type of valves, several conventional control strategies have been considered. For pneumatic actuator controlled by proportional directional control valve the usefulness of position control system using a fuzzy logic controller is designed and verified. In order to cope with a problem in fuzzy controller of exponential growth in rules with increase of the number of input variables, an adaptive fuzzy logic controller without a fuzzy rule base has been developed. For the sake of improving the accuracy and reducing the non-linear effects and time varying of parameters during the control process, a hybrid structure has been realized. This control concept is composed of PID controller tuned according to the damping optimum which is expanded with a non-linear friction compensator, and fuzzy logic controller designed in order to increase the robustness of the control system to variations of system parameters. The developed control algorithms are experimentally examined on a laboratory model of the pneumatic servo drive.