Abstract: Due to the lack of position information during manual ultrasonic non-destructive testing, it is hard to fully cover all inspection points and the defective locations have to be recorded manually. This paper presents a transducer positioning system using the combination of an optical mouse sensor and an inertial measurement unit. An adaptive Kalman filter is proposed to estimate the optimal angle of the transducer by fusing the data of acceleration and angular velocity measured from the inertial measurement unit. The position of the transducer is finally estimated by combining the optimal angle of the transducer and the displacement measured from the optical mouse. Experimental results show that the system can produce an accurate and robust position, which would be promising for automatically record the location of the transducer during manual ultrasonic inspection.