Snow and sea ice are two of the most important sources of water, which play significant roles in regional and global climate change and hydrological cycle. The MK test and lag analysis methods were employed to study the spatio-temporal characteristics of relationship between temperature and snow cover and relationship between temperature and sea ice based on NCEP reanalysis temperature data and NSIDC weekly snow and ice data. The results indicate that during the period of 1979-2013, the annual average temperature over the Northern Hemisphere and the Arctic showed a significant increasing trend, while the area of snow and sea ice showed a decreasing trend. Snow cover frequency decreased with the increase of temperature in most parts of the Northern Hemisphere, but increased with the increase of temperature in the lower Yangtze River Basin and Tibetan Plateau. In much of the near land sea, sea ice cover frequency significantly decreased with the increase of temperature. The variation of sea ice area was mainly attributed to the temperature change, and the highest negative correlation occurred when temperature change was 1-2 months ahead of the change of sea ice area. For snow cover, the temperature change is also the main reason. When the temperature change was 1-4 months ahead of the change of snow cover, the negative correlation index was the highest. Compared with snow, response of sea ice to the change of temperature is 1-2 months later, which was mainly caused by larger density of sea ice than that of snow. In general, the effect of temperature change on sea ice and snow cover has consistency with different response time to temperature and shows a strong spatial variability.