<p>Sensor failure is a common issue in pneumatic motion servo systems. To cope with the impact of pressure sensor failure on the system control performance, this study proposes a fault-tolerant control scheme based on a nonlinear pressure observer. When the pressure sensor fails, the observer replaces the sensor to monitor the air pressure inside the cylinder, thereby ensuring the stable operation of the system. Based on the model of the proportional valve-controlled cylinder pneumatic system, an integral sliding mode controller with a pressure observer is designed for the fault-tolerant control of the cylinder. To optimize the system control parameters, this study presents an improved sparrow search algorithm integrating osprey fishing and Gaussian mutation (OGSSA), and uses it for parameter tuning. Compared with several similar algorithms, the OGSSA exhibits superior convergence speed and accuracy on 22 out of 23 benchmark functions. Fault-tolerant control experiments are conducted based on the parameter combination optimized by the OGSSA. The results show that the designed pressure observer can effectively and quickly estimate the pressure inside the cylinder. Under three preset sudden fault conditions, the system does not show large transient errors, and can still stably track the motion trajectory. The increases in the maximum error and root-mean-square error are not greater than 10.2% and 5.2%, respectively. The results verify the effectiveness of using the pressure observer to solve the problem of pressure sensor failure.</p>