A common practice in biodiversity conservation is restoration of former species-rich grassland on ex-arable land. Major constraints for grassland restoration are high soil fertility and limited dispersal ability of plant species to target sites. Usually, studies focus on soil fertility or on methods to introduce plant seeds. However, the question is whether soil fertility reduction is always necessary for getting plant species established on target sites. In a three-year field experiment with ex-arable soil with intensive farming history, we tested single and combined effects of soil fertility reduction and sowing mid-successional plant species on plant community development and soil biological properties. A controlled microcosm study was performed to test short-term effects of soil fertility reduction measures on biomass production of mid-successional species. Soil fertility was manipulated by adding carbon (wood or straw) to incorporate plant-available nutrients into organic matter, or by removing nutrients through top soil removal (TSR). The sown species established successfully and their establishment was independent of carbon amendments. TSR reduced plant biomass, and effectively suppressed arable weeds, however, created a desert-like environment, inhibiting the effectiveness of sowing mid-successional plant species. Adding straw or wood resulted in short-term reduction of plant biomass, suggesting a temporal decrease in plant-available nutrients by microbial immobilisation. Straw and wood addition had little effects on soil biological properties, whereas TSR profoundly reduced numbers of bacteria, fungal biomass and nematode abundance. In conclusion, in ex-arable soils, on a short-term sowing is more effective for grassland restoration than strategies aiming at soil fertility reduction.