Strain monitoring is important for structural health monitoring of bridges, dams, high-rise buildings and so on. The metrological performance of strain sensing systems is an important guarantee for the accuracy of structural monitoring data. Common sensors for strain monitoring include Fiber Bragg grating strain sensors, resistive strain gauges, and vibrating wire strain gauges. Fibre Bragg grating strain sensors have the advantages of strong anti-electromagnetic interference, high precision, good durability, and distributed measurement. They have a large number of applications in the field of engineering construction and instrumentation. Fibre Bragg grating strain sensors usually work under dynamic conditions during bridge health monitoring. If the sensor measurement data is incorrect, it will have a negative impact on subsequent control, monitoring and fault diagnosis systems. Therefore, it is necessary to perform on-site calibration of the Fiber Bragg grating strain sensor in actual use to ensure the accuracy of the measurement results. At present, there are mainly two methods for on-site calibration of Fiber Bragg grating strain sensors: static calibration and dynamic calibration. In the actual bridge inspection, static calibration must block traffic operation for a period of time, which will have a certain impact on social and economic life, while dynamic calibration has little effect on traffic operation, so it is necessary to perform dynamic calibration of Fiber Bragg grating strain sensors. Aiming at the accuracy of fiber grating strain sensor in measuring dynamic strain value of bridge structure, a calibration method of Fibre Bragg grating strain sensor under dynamic excitation is proposed. The equal-strength cantilever beam is used to simulate the bridge structure, and weights of different weights are suspended instantaneously at the end of the equal-strength cantilever beam to simulate the dynamic excitation generated by the vehicle on the bridge. The resistance strain gauge has a simple structure, high output accuracy and good stability and is suitable for use as a standard sensor for the transmission of strain values. Therefore, a high-precision resistance strain gauge is selected as the reference sensor. The resistance strain gauge is used as the reference sensor, and the Fiber Bragg grating strain sensor is used as the sensor to be calibrated, and the measurement data sequences of the reference sensor and the sensor to be calibrated are compared. The premise of dynamic calibration is that the changes of the magnitude waveform of the sensor to be calibrated and the reference sensor under the same excitation source are basically the same. If the waveform of the two magnitude sequences is poorly matched, the dynamic calibration of the sensor cannot be performed. Because the sampling start time of the reference sensor and the sensor to be calibrated are not synchronized, and their own characteristics are different, the change waveform of the two measured value sequences will appear time misalignment. Therefore, in view of the time dislocation problem in the measurement data sequence, a cross-correlation algorithm is used to match the measurement data of the reference sensor and the sensor to be calibrated. Then, when the Fiber Bragg grating strain sensor has the initial value, the measurement value calibration is studied. When the Fiber Bragg grating strain sensor and the resistance strain gauge have the initial value, the ratio of the wave peak value of the two is used as the sensitivity calibration coefficient of the Fiber Bragg grating strain sensor. The experimental results show that the method effectively solves the time dislocation problem of the measurement data sequence, the data matching rate is more than 98%, and the dynamic calibration of the fiber grating strain sensor is realized. The use of different weights as the excitation source does not effect the calibration results. The sensitivity calibration coefficients of Fiber Bragg grating strain sensors at 1 529 nm and 1 547 nm wavelengths are basically consistent with the static calibration results.