The scattering cross section of Rayleigh-Brillouin is bigger than that of Raman scattering and it hence has an advantage in accurate tropospheric temperature profiling measurement. Moreover, accurate measurement of temperature under high pressure environment using Rayleigh Brillouin scattering is of great significance to monitoring of Space Shuttle Main Engine (SSME) Preburner and the scramjet engine. Both the deconvolution method and the convolution method are used to achieve the temperature retrieving of air under different pressures based on the spontaneous Rayleigh-Brillouin scattering. And the reasons induced the temperature retrival error are studied and a comparison of temperature measurement between the two methods is made. Before the deconvolution method based on Wiener filterbeing performed on the measured spectrum directly, the convolved spectra between the spontaneous Rayleigh-Brillouin scattering model and instrument transmission function are deconvolved to obtain the deconvolved spectra and the decovolved spectra are compared with the theoretical calculation spectra to retrieve temperatures. And the optimized singular value stacking number being 150, which is the key parameter of the deconvolution method, is obtained on account of temperature retriecal error being less than 1.0 K, the relatively unobvious fitting error and the short time consumption of retrieving temperature. And the spontaneous Rayleigh-Brillouin scattering spectra of air induced by the wavelength of 532nm of laser under the pressure of 1～7 bar at the temperature of 294.0 K are measured in experiment and the optimized scattering angle of 90.7° is obtained by the combination of theoretical spectrum and the principle of minimum value of χ２. After that, the deconvolution method and the convolution method are used to retrieve temperatures severally. Experiment results demonstrate that the spectral resolution is improved by using deconvolution method to some extent. Meanwhile, both the deconvolution method and the convolution method have good performance on temperature measurement under different pressures and the maximum errors between the retrieved temperature and the reference temperature are less than 2.0 K, temperature retrieving results of the deconvolution method are improved and time consumption of retrieving temperature is reduced with the pressure increasing, and temperature retrieving results using convolution method are better than those using the deconvolution method when the air pressure is low (≤2 bar), however, the results of both methods are close to each other and the absolute temperature errorsareless than 1. 0K when the air pressure is high (>2 bar). By analysis, it is found that the factors causing the temperature retrieval errors for both methods include the temperature fluctuations (±0.2 K), the effect of uncertainty of scattering angle and the known parameters on temperature retrieving and the spectral disturbancescaused by the nonlinear amplification of spectral noise of deconvolution method. The parameter measurement result can be improved in experiment by improving the signal-to-noise ratio of measured spectrum, the accuracy of optimized scattering angle and the deconvolution method.