Landscape pattern is the result of the interaction between human and nature. During the urbanization process, the landscape pattern had changed significantly, which affected the ecological process and thus threatened ecosystem services. The assessment of the impacts of landscape pattern on ecosystem services is crucial for regional sustainable development. Taking Beijing as a case study, we quantified the spatiotemporal changes of six selected ecosystem services, including food production, water purification, air quality regulation, carbon storage, habitat quality, and recreation opportunity from 1980 to 2018 by using the InVEST and Recreation Opportunity Spectrum (ROS) models. In addition, we analyzed changes of landscape patterns in Beijing, and explored the impacts of landscape patterns on ecosystem services. The results showed that:1) the landscape pattern in Beijing experienced dramatically changes from 1980 to 2018. The build-up land expanded by 120%, which mainly resulted in the reduction of cropland. The overall landscape pattern tended to be fragmented, the patch shape index became more complicated and the heterogeneity of the patches also increased. 2) Food production, water purification, carbon storage, habitat quality, and recreation opportunity showed a decreasing trend from 1980 to 2018. Spatially, food production was higher in the southeast and lower in the northwest. For the other ecosystem services, they were higher in the northwest, lower in the southeast, and the lowest in the central downtown area. During the past 38 years, the food production decreased the most in the central region of Beijing, while the air quality regulation, carbon storage, habitat quality, and recreational opportunity decreased the most in the eastern and southern regions. 3) At the landscape level, changes in land-use types, such as woodland, grassland, and build-up land had the most significant impacts on ecosystem services. Landscape metrics, such as the fragmentation index and diversity index influenced ecosystem services more obviously. At the class level, the mean patch area and the patch aggregation index had the most significant impacts on ecosystem services. Among them, the mean patch area and aggregation index of woodland showed positive effects on air quality and carbon storage. However, the aggregation index of built-up land showed a negative effect on habitat quality. In general, this study focused on analyzing the impacts of landscape pattern changes on ecosystem services, which can help decision makers to improve the ecosystem services by optimizing landscape patterns. The results of this study also can provide theoretical and practical basis for promoting regional landscape sustainable planning.