Electrocatalytic nitrogen oxide reduction (NOxRR) emerges as an effective way to bring the disrupted nitrogen cycle back into balance. However, efficient and selective NOxRR is still challenging partly due to the complex reaction mechanism, which is influenced by experimental conditions such as pH and electrode potential. Here, we have studied the enzyme-inspired iron single-atom catalysts (Fe-N4-C) and identified that the selectivity roots in the first step of the nitric oxide reduction. Combining the constrained molecular dynamics (MD) simulations with the quasi-equilibrium approximation, the effects of electrode potential and pH on the reaction free energy were considered explicitly and predicted quantitatively. Systematic heat maps for selectivity between single-N and N-N-coupled products in a wide pH-potential space are further developed, which have reproduced the experimental observations of NOxRR. The approach presented in this study allows for a realistic simulation of the electrocatalytic interfaces and a quantitative evaluation of interfacial effects. Our results in this study provide valuable and straightforward guidance for selective NOx reduction toward desired products by precisely designing the experimental conditions.