This paper presents a pneumatic muscle (PM) based active suspension control system which aims to effectively reduce vehicle vibration and acceleration so that the ride quality can be enhanced. Although the PM-based actuator inherits several advantages including power-weight ratio and convenient usage, the nonlinear and time-varying behaviors of PM cause lots of obstacles applied in the controller design. To control the PM-based actuator, this study adopts the robust control approach, with the adaptive scheme by combining the Haar wavelet series-based algorithm, to the design of the PM-based active vehicle suspension control system. The objective of using Haar wavelet series (HWS) function is to approximate an unknown function exists in the vehicle suspension system, the difficulty of model-based prerequisite can be bypassed to reduce the control design complexity. In addition, the embedded adaptive scheme is derived based on the Lyapunov approach and the overall system stability can be guaranteed. The validness of the proposed control system is demonstrated on the quarter-car suspension platform and the experimental results are presented.