Traditional mechanical water meters have gradually failed to meet the needs of modern society due to low accuracy, high maintenance costs, poor durability, and lack of digital management capabilities. In order to improve the measurement accuracy and adaptability of the water meter, this study uses the ultrasonic water meter, based on the time difference method, and proposes an improved Kalman Filter Algorithm for the measurement of the ultrasonic water meter. In this paper, the measurement principle of an ultrasonic water meter is described in detail. The measurement accuracy under different flow rates and temperature conditions is verified by experiments. The experimental results show that the improved Kalman Filter Algorithm can effectively reduce the influence of noise. Even under conditions of low flow rate and large temperature change, the measurement error can be controlled within 1 %, which improves the measurement accuracy of the ultrasonic water meter.