为提高分布式光纤拉曼测温系统的测量速度和测量准确度，提出了一种自补偿光纤损耗及光纤色散的温度解调方法，并进行了实验验证。该方法对斯托克斯与反斯托克斯后向散射信号进行了损耗修正，避免了测温前对整条传感光纤进行定标处理的过程，减小了系统的运行时间；采用色散补偿平移算法对斯托克斯后向散射信号的位置进行修正，获得了与反斯托克斯后向散射信号相同位置处的斯托克斯后向散射信号的强度，降低了光纤色散对温度解调的影响，提高了系统的测温准确度。实验结果表明，当光纤传感距离为5.8 km时，温度波动由9.01 ℃下降到0.57 ℃，测温准确度由5.50 ℃优化至0.87 ℃。

To improve the measuring speed and the measuring accuracy of the distributed fiber Raman temperature measurement system, a temperature demodulation method of self-compensation for fiber loss and fiber dispersion is proposed, and the experimental verification is carried out. The loss between Stokes backscattering signal and anti-Stokes backscattering signal is modified by the proposed method, the calibration processing for the entire sensing fiber before temperature measurement is avoided, and the system operating time is reduced. The dispersion compensation shift algorithm is applied to correct the position of the Stokes backscattering signal. The intensity of Stokes backscattering signal can be obtained at the same position of the anti-Stokes backscattering signal, which eliminates the influence of fiber dispersion on temperature demodulation and improves the accuracy of temperature measurement. The experimental results indicate that the temperature fluctuation is declined from 9.01 ℃ to 0.57 ℃ when the fiber sensing distance is 5.8 km, and the temperature accuracy is optimized from 5.50 ℃ to 0.87 ℃.