In order to further understand the effects of climate on forest-runoff relationship, the experimental results from different catchments are employed to detect spatial variation in forest-runoff relationship. Based on the numerical simulations of water balance of the Suomohe Basin (a tributary of the Yangtze River in the mountain region of southwestern China) which are conducted under different temperature change scenarios using the coupled model (Biophysical/Dynamic Vegetation Model SSiB4/TRIFFID coupled with TOPMODEL), the effects of water and heat factor on forest-runoff relationship are analyzed through analyzing the impacts of temperature and soil water on net canopy photosynthesis of forests, canopy wetness fraction, transpiration, canopy interception evaporation, and soil evaporation. The main conclusions are as follows: water and heat conditions affected forest-runoff relationship by controlling the net photosynthetic rate, canopy wetness fraction, transpiration, canopy interception evaporation, and soil evaporation. With the increase of temperature, the increase of forest evapotranspiration is more than that of other vegetation types and bare soil, and leading to changes in forest-runoff relationship. The spatial variation of water and heat conditions determines the spatial variation of water use efficiency and forest-runoff relationship. As temperature increases duo to the vertical and the horizontal climate zones, water use efficiency decreases and forest-runoff relationship exists from increasing runoff to no significant effect on runoff and reducing runoff. Water stress or high temperature will inhibit forest transpiration and reduce the role of forests in reducing runoff. Results from catchment experiments also indicate such spatial variation in forest-runoff relationship.