This research develops a multiple-surface sliding mode control (MSSC) approach for position control of a servo-pneumatic system in the presence of mismatched uncertainties due to the friction force of the cylinder sealing. Servo-pneumatic actuators have many applications such as industrial automation, haptic interfaces, rehabilitation robots and non-invasive surgeries. Non-linearities due to internal and external disturbances, such as the friction force between the piston seal and cylinder wall, make it difficult to achieve adequate performance from these actuators. Thus, modelling and identification of friction parameters is an essential part of the controller design procedure. A simple model for friction such as the Stribeck model may be used in order to reduce the complexity of the identification procedure. In addition, a bounded uncertainty owing to the unmodelled dynamics of friction is considered. The lack of direct measurement for systems velocity necessitates finding a solution to estimate this parameter. One answer to this problem is to employ a high-gain observer. Whereas mismatched uncertainties appear in the state space equation of the system before control input, control input cannot apply directly to them. In the proposed framework, an MSSC scheme has been used to cope with these types of uncertainties. Asymptotic stability of the closed-loop system is proven by using the Lyapunov method; experimental results show that the proposed controller can deliver a good tracking performance and is robust to uncertainties. Experimental results validate the controller performance.