The current artillery chamber temperature test has problems such as high testing difficulty,low accuracy and harsh conditions.In this paper,a fire chamber temperature measuring system based on colorimetric temperature measurement is designed to solve above problems.It is proved that the selection of wavelength and calibration error are two important factors affecting the accuracy of colorimetric thermometry.By analyzing the sensitivity relationship between wavelength and acquisition voltage,the optimal band is 0.960μmand 1.000μm,which improves the accuracy of colorimetric thermometry.And the least square method and genetic algorithm(GA)are used to fit the experimental data so as to test and verify the system.The results show that choosing the optimal wavelength can improve the accuracy of temperature measurement.Compared with the least square method,the accuracy of genetic algorithm is significantly improved,and the overall measurement error is less than 4 ‰.