In this article, we propose a novel method for real-time detection of NaCl concentration based on ultrasonic impedance method. Specifically, our method is based on the fact that different concentration of NaCl leads to differences in acoustic characteristics of ultrasonic waves. We conducted an experimental study on the relationship between the concentration and the amplitude of ultrasonic echo. Our evaluations show that our approach can push the linear fitting degree r2 to 0.9967. Then we used the BP neural network to classify and identify the collected sample data and achieved 100% classification accuracy. We also established a concentration-amplitude prediction model based on the BP neural network, and the result shows that the error between the predicted value and the ground truth is less than 8.36%. This method realizes non-contact measurement and can be applied to real-time monitoring of NaCl concentration in large, closed containers.