Land-use legacies are important for explaining present-day ecological patterns and processes. However, an overarching approach to quantify land-use history effects on ecosystem properties is lacking, mainly due to the scarcity of high-quality, complete and detailed data on past land use. We propose a general framework for quantifying the effects of land-use history on ecosystem properties, which is applicable (i) to different ecological processes in various ecosystem types and across trophic levels; and (ii) when historical data are incomplete or of variable quality. The conceptual foundation of our framework is that past land use affects current (and future) ecosystem properties through altering the past values of resources and conditions that are the driving variables of ecosystem responses. We describe and illustrate how Markov chains can be applied to derive past time series of driving variables, and how these time series can be used to improve our understanding of present-day ecosystem properties. We present our framework in a stepwise manner, elucidating its general nature. We illustrate its application through a case study on the importance of past light levels for the contemporary understorey composition of temperate deciduous forest. We found that the understorey shows legacies of past forest management: high past light availability lead to a low proportion of typical forest species in the understorey. Our framework can be a useful tool for quantifying the effect of past land use on ecological patterns and processes and enhancing our understanding of ecosystem dynamics by including legacy effects which have often been ignored.