Pneumatic cylinder actuators are significantly utilized for industry automatic systems in the fields of mechanical applications. We propose a novel control method for pneumatic cylinder actuator systems including stochastic friction dynamics. The proposed control mechanism is linearly composed of nominal control and auxiliary control variables. The former is designed from linear system model without friction terms by using a previous linear system theory and the latter is constructed as a function of friction estimation which is carried out by a well-known least square algorithm for reducing the control error due to random friction dynamics. We accomplish numerical simulation to demonstrate reliability of the proposed control method and conduct a comparative study to improve its superiority.