The traditional method of calculating the height of a measurement point through atmospheric pressure and standard atmospheric models ignores the irregular changes in the atmosphere, resulting in significant errors. In this paper, a mathematical model for dynamic differential measurement is derived on the basis of the basic atmospheric pressure-height theoretical model. Then, a high-precision elevation sensor based on atmospheric pressure is developed using IoT technology, and a real-time dynamic differential measurement system is established. By establishing a ground reference station to dynamically calibrate the errors caused by atmospheric changes in real time, high-precision and highly reliable real-time elevations are obtained through calculation. The experiment shows that the average measurement error due to dynamic differences is less than 5 cm. At the same time, we also use the sliding average filtering method to effectively improve the real-time measurement accuracy while ensuring the constant measurement frequency. The measurement error is reduced from ± 15 to ± 5 cm, and the measurement results can be uploaded to a cloud platform in real time, with good engineering practicality.