The impact of agricultural practices on CO2 emissions from soils needs to be understood and quantified to enhance ecosystem functions, especially the ability of soils to sequester atmospheric carbon (C), while enhancing food and biomass production. The objective of this study was to assess CO2 emissions in the soil surface following tillage abandonment and to investigate some of the underlying soil physical, chemical and biological controls. Maize (Zea mays) was planted under conventional tillage (T) and no-tillage (NT), both without crop residues under smallholder farming conditions in Potshini, South Africa. Intact top-soil (0–0.05 m) core samples (N = 54) from three 5 × 15 m2 plots per treatment were collected two years after conversion of T to NT to evaluate the short-term CO2 emissions. Depending on the treatment, cores were left intact, compacted by 5 and 10%, or had surface crusts removed. They were incubated for 20 days with measurements of CO2 fluxes twice a day during the first three days and once a day thereafter. Soil organic C (SOC) content, soil bulk density (ρb), aggregate stability, soil organic matter quality, and microbial biomass and its activity were evaluated at the onset of the incubation. CO2 emissions were 22% lower under NT compared with T with CO2 emissions of 0.9 ± 0.10 vs 1.1 ± 0.10 mg C–CO2 gC−1 day−1 under NT and T, respectively, suggesting greater SOC protection under NT. However, there were greater total CO2 emissions per unit of surface by 9% under NT compared to T (1.15 ± 0.03 vs 1.05 ± 0.04 g C–CO2 m−2 day−1). SOC protection significantly increased with the increase in soil bulk density (r = 0.89) and aggregate stability (from 1.7 ± 0.25 mm to 2.3 ± 0.31, r = 0.50), and to the decrease in microbial biomass and its activity (r = −0.59 and −0.57, respectively). In contrast, the greater NT CO2 emissions per m2 were explained by top-soil enrichment in SOC by 48% (from 12.4 ± 0.2 to 19.1 ± 0.4 g kg−1, r = 0.59). These results on the soil controls of tillage impact on CO2 emissions are expected to inform on the required shifts in agricultural practices for enhancing C sequestration in soils. In the context of the study, any mechanism favoring aggregate stability and promoting SOC allocation deep in the soil profile rather than in the top-soil would greatly diminish soil CO2 outputs and thus stimulate C sequestration.