Abstract  The relationship and feedback between landscape pattern, function and process serve to describe the behavior of a regional landscape. Based on landscape function characteristics such as biological productivity, soil nutrient content, vegetative cover, etc., a quantitative method and digital model for analyzing evolving landscape functionality in the headwaters area of the Yellow River in the People's Republic of China were devised. Through the analysis of three-phase remote sensing data from 1975, 1985 and 1995 and based upon the well-defined characteristics of this region's evolving landscape over the past 30 years, the attendant ecology of the different functional landscape ecotypes was investigated. Between 1975 and 1995 the area of AC&S (alpine cold meadow and steppe) in the source area of the Yellow River has decreased by 27.25%, ACSW (alpine cold swamp meadow) has decreased by 27.04%, ALP (alpine steppe) by 38.18% and lakes by 9.78%. The grass biomass production decreased by 752.37 Gg, of which AC&S meadows accounted for 83.8% of these losses. The overall stock capacity of the headwaters area of the Yellow River decreased by 518.36 thousand sheep units. Soil nutrients showed a similar pattern, soil nutrient loss was greater from 1985¨C1995 than from 1975¨C1985. Changes in the overall ecological functionality of the area were not simply a result of a summation of the changes associated with individual evolving landscapes, but rather an integration of positive and negative influences. Landscape evolution occurs in two main directions: degradation and strengthening (expanding and improving). An understanding of the direction, force and integration of parameters influencing landscape evolution as it impacts the attending ecosystems can allow one to foresee how the landscape of the Yellow River source area will evolve in the coming years.