Measurements strategies based on absorption spectroscopy techniques, especially the measurements in high temperature, require accurate values of important spectroscopic parameters of the probed species. Sometimes the parameters listed in widely used HITRAN and HITEMP2004 database are uncertain to some extent. In order to validate the spectroscopic parameters of 9 selected CO2 lines which should be used in combustion diagnosis, spectra of those lines were recorded in a high temperature experiment setup as a function of temperature (in the range of 300-800 K) and pressure (in the range of 9-450 torr) using a distributed feed-back (DFB) diode laser. The recorded absorption spectra were fitted to Voigt profile. Line intensity, air-broadening coefficient and temperature exponent of each line were deduced from those data. Through comparison of experimental results and those listed in HITRAN and HITEMP2004 database, the discrepancies of most line intensities, air-broadening coefficients and their temperature exponents are less than 3%, 5% and 2% respectively. Those results show good consistency between the experimental data and that in HITRAN and HITEMP2004 database. The discrepancy in line intensities may be caused by the fitting of absorption spectra, the reading of thermocouple and pressure gage, uniformity of temperature in the heated cell, and uncertainty of the optical path. Those factors also cause the discrepancy in air-broadening coefficients and their temperature exponent. CO2 contained in air also introduces error in air-broadening coefficients and their temperature exponent beside those factors. Though we have deducted them in data-processing, the little change of CO2 in partial region also exists. Those results will be helpful to the measurement of CO2 concentration in combustion diagnosis in the future.