Soil water is known to be a key factor for controlling N2O emissions. The relationship between N2O emissions and water-filled pore space (WFPS) at hydric equilibrium is often described as exponential, starting at 60% WFPS. Because water in soil is dynamic, however, N2O emissions should be seen as a dynamic process. In this study, we investigated the role of the soil hydric history on N2O emissions by varying the intensity and speed of soil drying. We performed a laboratory experiment under non-steady-state equilibrium to control the bottom boundary condition of undisturbed soil cylinders. Three treatments of stepwise drying were performed across the range 0 to −100 cm pressure head in two wetting-drying cycles. Two types of N2O peaks of the same order of magnitude were detected. A peak appeared within 2 d after rewetting and was related to microbial production processes. The second type of peak was detected within the first two steps of the decreasing pressure head. They occurred during a brief period, an average of 1.6 h after the change in hydric equilibrium, during the phase of steep decline in the soil matric potential. These peaks were induced by diffusion processes, and their intensities were correlated with the amount of water drained. Entrapment of N2O during the wetting phase and rapid displacement during the drying phase may have occurred. Our results showed that WFPS alone is not able to consistently predict N2O emissions by soils because production and emission phases must be distinguished.