This paper is concerned with the fixed-time consensus problem of multiple chained-form systems under matched perturbations. In this study, the leader (which can be dynamic) only transmits its state and control input to its neighbors. For each agent, a decentralized observer is designed to estimate the leader state in a fixed-time. Contrary to finite-time schemes, the estimation of the settling time does not require the knowledge of the initial state, allowing a step-by-step design for the controller. A decentralized observer-based control protocol is proposed for each agent to solve the leader–follower consensus problem in a fixed-time. This paper ends with a numerical example showing the effectiveness of the proposed approach.