Although continuous no-till (NT) is recommended for erosion control and carbon sequestration, it often has a limited duration since farmers alternate between NT and full inversion tillage (FIT) to control weed infestation and avoid soil compaction. In this paper, we evaluate the effect of continuous tillage and tillage conversion of NT to FIT and vice versa on SOC and SON stocks, in a long-term experiment at Boigneville in Northern France. Continuous NT (CNT) and FIT (CFIT) treatments were established in 1991 and maintained until 2011 while half of the plots were converted in 2005: from CNT to new FIT (NFIT) and CFIT to new NT (NNT). Bulk densities and organic C and N contents were determined in 2001 and 2011 down to the old ploughing depth (opd) which was also measured. SOC and SON stocks were calculated at equivalent soil mass by correcting either bulk densities or the opd. Both methods produced very close results and similar conclusions. A typical gradient of SOC and SON concentrations vs depth was observed in CNT as opposed to a rather uniform distribution in CFIT. CNT resulted in SOC concentration in the top soil (0–5 cm) higher by 38% in 2001 and 53% in 2011 compared to CFIT. Conversely, it led to a SOC reduction in the deeper layer (ca. 10–28 cm) by 14% in 2001 and 18% in 2011. The global effect was no significant change in SOC and SON stocks between treatments over the old ploughed layer (4060 t soil ha−1) in both years: 43.2 and 45.0 t C ha−1 in 2001 and 44.7 and 45.8 t C ha−1 in 2011, in CNT and CFIT, respectively. In 2011, six years after tillage conversion, the stratification of SOC and SON had disappeared in NFIT whereas a new one had appeared in NNT with a smaller gradient than in CNT. SOC or SON stocks over the old ploughed layer did not differ significantly between treatments after 6 years of conversion: SOC stocks were 45.8, 43.2, 44.7 and 43.1 t C ha−1 in the CFIT, NFIT, CNT and NNT treatments, respectively. Furthermore, SOC stocks below the old ploughed layer (ca. 28–40 cm) were slightly greater in FIT than in NT treatment (10.9 vs 8.7 t C ha−1). In this experiment, continuous or conversion tillage did not result in any C sequestration benefit.