Land-use intensification has contrasting effects on different ecosystem services, often leading to land-use conflicts. Multiple studies, especially within the 'land-sharing versus land-sparing' debate, have demonstrated how landscape-scale strategies can minimise the trade-off between agricultural production and biodiversity conservation. However, little is known about which land-use strategies maximise the landscape-level supply of multiple ecosystem services (landscape multifunctionality), a common goal of stakeholder communities. Here, we combine data collected from 150 grassland sites with a simulation approach to identify landscape compositions, with differing proportions of low-, medium-, and high-intensity grasslands, that minimise trade-offs between the four main grassland ecosystem services demanded by stakeholders: biodiversity conservation, aesthetic value, productivity and carbon storage. We show that optimisation becomes increasingly difficult as more services are considered, due to varying responses of individual services to land-use intensity and the confounding effects of other environmental drivers. Thus, our results show that simple land-use strategies cannot deliver high levels of all services, making hard choices inevitable when there are trade-offs between multiple services. However, if moderate service levels are deemed acceptable, then strategies similar to the 'land-sparing' approach can deliver landscape multifunctionality. Given the sensitivity of our results on these factors we provide an online tool that identifies strategies based on user-defined demand for each service (https://neyret.shinyapps.io/landscape_composition_for_multifunctionality/). Such a tool can aid informed decision making and allow for the roles of stakeholder demands and biophysical trade-offs to be understood by scientists and practitioners alike.