To realize the on-line measurement of the ion tube’s Lissajous figure. This article uses BP neural network to fit the nonlinear relationship between the voltage and current signal of low-voltage side and the total voltage and measurement voltage signal of high-voltage side of the ion tube transformer. A data acquisition circuit is designed to obtain the sample data used for training and testing the BP neural network, and the test result shows that the computational error of the constructed BP neural network is less than 3%. In the actual measurement process, in order to get the Lissajous figure of ion tube, the high-voltage signal is calculated from the low-voltage signal of the ion tube transformer by using the well-trained BP neural network. Finally, based on the STM32 microcontroller, an on-line measurement system prototype of ion tube Lissajous figure is designed. The experimental analysis shows that the designed BP neural network is suitable for calculating the Lissajous figure for the same ion tube with different working voltages and different ion tubes with the same working voltage, and the designed on-line measurement system prototype can realize the on-line measurement of ion tube’s Lissajous figure.