Adjusting fertility programs according to peak demand periods will help prevent periodic nutrient disorders during crop growth, allowing enhanced use efficiency of water and fertilization. The objectives of this article were to investigate 1) the evolution of the changes in the concentration of nitrate and ammonium in the recycled solution with different N-NO3
–/N-NH4
+ ratios; 2) the influence of the N form supply (N-NO3
– or N-NH4
+) on the nitrogen uptake, the nitrogen:water uptake ratio, and nitrogen use and uptake efficiency; and 3) the development of empirical models that would allow the prediction of nitrogen nutritional needs of Dieffenbachia amoena to increase the N use efficiency in a recycled system. To achieve these aims, N uptake has been correlated to climate parameters such as temperature, vapor pressure deficit and global radiation, and growth parameters such as leaf area index. The trial was carried out with Dieffenbachia amoena plants growing in a recycled system with expanded clay as substrate. The crop was placed in an INSOLE (Buried Solar Greenhouse), the plants being supplied with equal amounts of N, differing in the percentage of the N form applied (NO3
–:NH4
+): TA (100:0), TB (50:50), and TC (0:100). The nitrogen form generated important changes in the pH and nitrate and ammonium concentration in nutritive solution during the recirculating solution. In N-NO3
– treatment, pH increased and nitrate concentration showed a tendency to drop slightly. N-NH4
+ treatment showed an abrupt N-NH4
+ concentration decrease, and N-NO3
– concentration increased along with a pH drop. Also, the nitrogen form applied to the Dieffenbachia amoena crop affects nitrogen uptake. Nitrogen uptake rates and nitrogen uptake concentration were higher in the plants supplied with N-NH4
+ or NO3
–+NH4
+ than in the plants provided with N-NO3
– alone. The supply of a combination of 50:50 NO3
–+NH4
+ improved the N use efficiency. The study also indicated the possibility of predicting the N uptake rate and N uptake concentration using the proposed models.