The delivery of multiple ecosystem services from intensively managed cropping systems remains challenging mainly because increases in crop yields are rarely associated with greater soil carbon (C) sequestration or efficient ecosystem nitrogen (N) cycling. The sustainability of intensive cropping systems depends on trade-offs between multiple ecosystem services but experimental evidence from long-term field studies remains limited. The common expectation is that highly productive agroecosystems will be associated with lower soil C and reduced nutrient use efficiency. Here we use data spanning 30 years of a long-term field experiment established in 1985 near Lodi in Northern Italy where four arable systems (i.e., three crop rotations and a grain maize monoculture) and a no-till permanent meadow were all compared across two levels of agronomic inputs. We asked how (i) soil C stocks, (ii) fertilizer N-use efficiency, (iii) productivity, (iv) crude protein, and (v) feed units for lactation might differ across the five traditional cropping systems. We found that soil C sequestration, N-use efficiency, and crude protein production were all significantly higher in the no-till permanent meadow, which was however associated with lower dry matter yields and reduced feed units for lactation when compared with the arable systems. These findings suggest that converting arable soil into no-till permanent meadow is a win-lose solution. The most productive annual rotation (silage maize + Italian ryegrass) offers a complementary win-lose solution whereby high yields and feed units for lactation can be obtained from a smaller cropland area. The other cropping systems are less productive and provide little (or none) soil C storage benefits. Our study is one of the first to show clear trade-offs between multiple ecosystem services and to demonstrate that high fertilization rates might be only justified when the management goal is to spare land for less intensive uses such as permanent meadow.