Studies about nitrogen (N) mineralization and nitrification in deep soil layers are rare because N processes are considered to occur mainly in topsoil that hosts active and diverse microbial communities. This study aimed to measure the soil potential net N mineralization (PNM) and nitrification (PNN) down to 4 m depth and to discuss factors controlling their variability. Twenty-one soil cores were collected at the Restinclières agroforestry experimental site, where 14-year-old hybrid walnut trees were intercropped with durum wheat. Soil cores were incubated in the dark in the laboratory at both 6 and 25°C. The soil was a deep calcic fluvisol with a fluctuating water table. It featured a black layer that was very rich in organic matter and permanently water saturated at depths between 3.0 and 4.0 m. The mean soil mineral N content was 3 mg N kg−1 soil in the upper 0.0–0.2 m layer, decreasing until a depth of 2 m and increasing to the maximum value of 25.8 mg N kg−1 soil in the black layer. While nitrate (NO3−) was the dominant form of mineral N (89%) in the upper 0.0–0.2 m layer, its proportion progressively decreased with depth until ammonium (NH4+) became almost the only form of mineral N (97%) in the saturated black layer. Laboratory soil incubation revealed that PNM and PNN occurred at all depths, although the latter remained low at 6°C. The soil nitrate content in the black layer was multiplied by 48 times after 51 days of incubation at 25°C, whereas it was almost inexistent at the sampling date. While the soil total N, the pH and the incubation temperature explained 84% of the variation in PNM, only 29% of the percent nitrification variance was explained by the incubation temperature (Tinc) and the soil C-to-N ratio. These results point out the necessity to consider soil potential net N mineralization and nitrification of deep soil layers to improve model predictions.