|This paper proposes a fault detection algorithm for a velocity controller using MRAC in the belt conveyor system of a fish processing line. In general terms, the fault detection algorithm utilizes an extended Kalman filter (EKF) based on a known mathematical model to estimate the state vector and compare computed residue values with a threshold value to determine fault condition. However, a system that models a belt conveyor system has some uncertain parameters. Thus, a fault detection algorithm based on the mathematical model with uncertain parameters can give false alarms or miss detection events. The following can be done to solve this issue: First, a velocity controller based on an MRAC has been designed to estimate uncertain parameters in the belt conveyor system and control belt conveyor velocity to track a reference input asymptotically. Second, an angular velocity output and a friction coefficient are chosen as the state vector and observation vector elements for the EKF. Third, process and observation vector functions are determined and the EKF is used to compute the residue value, which is then compared to a given threshold value to detect a fault. Finally, the simulation and experimental results are shown to verify the proposed algorithm’s effectiveness.