The present study evaluates the relationship between the crops productivity and ecosystem diversity. The spatial variability in ecosystem diversity was measured using the Shannon landscape diversity index and distance from biodiversity hotspots that are nature conservation areas. Three crops were selected for the study: soybeans, sunflowers and winter rye. The initial data included the average crops yields in administrative districts within 10 regions of Ukraine. It was found that the studied crops yield dynamics from the mid-90s of the previous century to the current period could be described by a sigmoid curve (log-logistic model). The parameters of the yield model are the following indicators: the minimum level of yield (Lower Limit); maximum level of productivity (Upper limit); the slope of the model, which shows the rate of change in yields over time; ED50 - the time required to achieve half, from the maximum yield level. Our studies have shown that there is a statistically significant regression relationship between the yield parameters of all the studied crops and biodiversity, even at the landscape level. Among the studied crops, soybean shows the strongest regression relationship between yields and indicators of landscape diversity. Sunflower yield is the least dependent on landscape diversity. Most of the established dependencies are nonlinear, which indicates the existence of an optimal level of landscape diversity to achieve the maximum possible crop yields. Therefore, the obtained patterns can be the basis for land-use planning and management, especially while creating new natural protected areas.