In the magnetic confinement device, the existence of impurity, especially the high-Z impurity, will cause the large enhancement of the radiation power loss and serious degradation of plasma performance. Quantitative study of impurity behavior requires the absolute intensity calibration of spectroscopy diagnostic system for impurity firstly. In this work, a precise in-situ absolute intensity calibration of a fast-response flat-field extreme ultraviolet (EUV) spectrometer is carried out by using different methods in different wavelength range. In the range of 20～150 , the sensitivity curve is obtained by comparing EUV bremsstrahlung between the measurement and calculation, in which the latter one is calculated by combining the profile of electron density, temperature and the effective charge Zeff deduced from the absolutely measured visible bremsstrahlung intensity in the range of (523±1) nm. The measured EUV bremsstrahlung intensity is the continuous counts subtracting the background noise at different wavelengths from the detector. In the longer wavelength range, i. e. 130～280 , the relative intensity calibration is addressed by comparing the measured and simulated line intensity ratio of resonance transition doubles from Li-like ions, e. g. Fe23+, Cr21+, Ar15+, and Na-like ions, e. g. Mo31+, Fe15+, with the transition of 1s22s2S1/2 —1s22p2P1/2, 3/2 and 2p63s2S1/2—2p63p2P1/2, 3/2, respectively. For the simulated line intensity ratio, it is modeled using the collisional-radiative model, in which the energy level populations are determined by electron collisional excitation, de-excitation and radiative decay. By combining those two methods, the absolute calibration of EUV spectrometer is achieved in the wavelength range of 20～280 . In addition, the uncertainty of the calibration is estimated to be about 30%, considering the measurement uncertainties of electron temperature, electron density and bremsstrahlung. Based on the obtained absolute sensitivity curve, the quantitative study of impurity concentration is being carried out by comparing the absolute measurement of emission line intensity with that simulated by combining the one-dimensional impurity transport simulation and ADAS (Atomic Data and Analysis Structure) database.