The identification of ecosystem service trade-off characteristics and drivers of cropland is an important basis for implementing scientific governance decisions for cropland resources. This paper constructs an assessment framework for cropland ecosystem services, and analyzes the trade-off/synergy relationship, dominant drivers and characteristics of cropland ecosystem services in the Yangtze River Delta region (YRD) from 2000 to 2020 by using the InVEST model, Spearman's correlation analysis, Root-mean-square error, and GBDT methods. The results show that (1) from 2000 to 2020, grain production in the YRD continued to improve, with the core area distributed in the central and northern plains. Water yield service and carbon sequestration and oxygen release both show a decreasing and then increasing trend, with a spatial distribution of “high in the south and low in the north, high in the east and low in the west”. The amount of soil retention continues to decline, with high values concentrated in the forest ecological barrier area in the southwestern part of the country. Cultural and recreational service showed a continuous growth, and the distribution of high values changed to a pattern of “large concentration and small dispersion”, and the ecosystem services of cropland showed the direction of development of leisure and living. (2) The trade-off/synergy relationship of cropland ecosystem services in the YRD is characterized by obvious spatial and temporal heterogeneity. In time, it is generally characterized by the weakening of trade-off and the strengthening of synergy, and the trade-off/synergy relationship will change in intensity or even reverse in different periods. Cropland in the YRD brings ecological risks while improving agricultural efficiency. With the transformation of cropland use and the development of agricultural modernization, cropland shows a trend of comprehensive optimization of food production, ecological value, culture and recreation. Spatially, due to different natural conditions and economic levels, the production capacity of cropland and the level of factor input vary widely at the spatial level, so that the overall trade-offs between services are manifested in two distribution patterns: high in the north and south, low in the middle, or high in the south and low in the north. The overall improvement in the technology and capital inputs of cropland production has contributed to a further narrowing of the intensity of trade-offs across the region, and the characteristics of cropland development in different regions have gradually come to the fore. (3) The contribution of cropland resource endowment to the trade-off intensity declined in 2000—2020, and the contribution of socio-economic factors increased substantially, with light conditions, agricultural policies and urbanization level being the key factors affecting the trade-off intensity of cropland ecosystem services, showing a non-linear change characteristic of the existence of a threshold effect. With the passage of time, the threshold range of light duration generally remained unchanged at 1800—2000h, and the thresholds of cropland replanting index and urbanization rate showed an increasing trend, in which the cropping intensity increased from 1.3—1.6 to the range of more than 1.6, and the urbanization rate from 40%—60% The trade-offs among ecosystem services overall weakened when the threshold was raised to near 80%. In order to promote synergistic development among cropland ecosystem services, it is necessary to control the driving factors within a certain threshold range.