Cropland abandonment is an opportunity for semi-natural biodiverse ecosystems to develop through spontaneous succession or active restoration. Spontaneous recovery is limited by the dispersal and capacity of species to establish under the given environmental circumstances. This paper provides an evaluation of three restoration techniques to overcome dispersal and environmental limitations based on revisiting an experiment with the aim to restore Pannonian sand grasslands on old fields, 16 years after its start. Treatments were carried out in two sites with different date of abandonment, initial soil conditions and landscape features. The experimental design involved ploughing as pm-treatment and eight types of treatments, which consisted of the combinations of initial seeding (2002/2003), mowing (2003-2008), carbon amendment (2003-2008) and control in eight replicates per site. We compared the impact of seeding, mowing, carbon amendment and site on soil available nitrogen, species richness, total vascular plant cover and the relative cover of target and neophyte species with the use of GLM and one-way and two-way ANOVA. We also analysed site-related factors such as soil characteristics and landscape context. We found a long-term impact of mowing on the total vascular plant cover. We also detected an impact of seeding on the relative cover of seeded species, mainly due to Stipa borysthenica (18.5% in seeded and 8.5% in non-seeded plots), and on the relative cover of neophyte species, in particular the annual neophyte, Dnbrosia artemisiifolia (0.07% in seeded and 2.32% in non-seeded plots in one site). Seeded species established and persisted in the long term and also spread into non-seeded plots resulting in 85% and 91% mean cover of target species. Most of the parameters analysed, with the exception of the cover of Festirca vaginata, demonstrated a site-specific vegetation development. No long-term impact on soil-available nitrogen was found. The results of our long-term evaluation of restorative treatments indicate that seeding can successfully assist the restoration of sand grasslands on abandoned fields, while carbon amendment can be used to supplement seeding in the initial years and infrequent mowing in the later years of restoration. The spread of seeded species into non-seeded plots did not allow for the causal analysis of single factors such as mowing or carbon addition in the long term, but it has optimised the restoration process. Based on our results we suggest that restoration of old fields can be triggered with the use of small seed introduction units, also referred to as establishment windows.