As an important underlying surface, the thermal properties of snow are extremely important to the surrounding environment. Based on the snow temperature data under different densities (0.30g/cm3, 0.35g/cm3, 0.40g/cm3 and 0.45g/cm3) of natural snow in Northeast China, the variation characteristics of snow temperature were analyzed. At the same time, the variation law of the thermal conductivity of snow at different temperatures (-5℃, -10℃, -15℃, -20℃ and -25℃) and densities (0.35g/cm3, 0.40g/cm3, 0.45g/cm3, 0.50g/cm3, 0.55g/cm3 and 0.60g/cm3) was analyzed. The results showed that the thermal conductivity of snow was increased with the increase of density, and there was a great index correlation between them. And the response of the thermal conductivity of snow in temperature showed obvious difference under different densities. In the density range of 0.40~0.45g/cm3, the thermal conductivity was increased with the increase of temperature;in the density range of 0.50~0.60g/cm3 and the temperature range of -15~-25℃, the thermal conductivity of snow was increased with the decrease of temperature;during the temperature observation period, snow cover temperature was consistent with the change trend of air temperature, and both were less than 0℃. In different periods of the day, the surface of the snow layer appeared to the cold and warm center, and the temperature of the shallow 0~20cm snow layer was changed dramatically. The variation range of snow temperature was gradually decreased with the increase of snow depth. There was a great negative index correlation between amplitude of snow temperature and snow depth. The time of reaching the extreme value of deep snow temperature was slightly lagged behind that of the surface snow layer about 4h. In the same depth of the snow layer, the temperature was decreased with the increase of density.