Many researchers have investigated pneumatic servo positioning systems due to their numerous advantages: inexpensive, clean, safe, and high ratio of power to weight. However, the compressibility of the working medium, air, and the inherent nonlinearity of the system continue to make achieving accurate position control a challenging problem. In this paper, two control algorithms are designed for the position tracking problem and their experimental performance is compared for a pneumatic cylinder actuator. The first algorithm is sliding-mode control based on a linearized plant model (SMCL) and the second is sliding-mode control based on a nonlinear plant model (SMCN). Extensive experiments using different payloads (L9, 5.8, and 10.8 kg), vertical and horizontal movements, and move sizes from 3 to 250 mm were conducted. Averaged over 70 experiments with various operating conditions, the tracking error for SMCN was 18% less than with SMCL. For a 5.8-kg payload and a 0.5-Hz; 70-mm amplitude, sine wave reference trajectory, the root-mean-square error with SMCN was less than 0.4 nun for both vertical and horizontal motions. This tracking control performance is better than those previously reported for similar systems.