The nitrogen/N dynamic is complex and affected by soil management (i.e., residue accumulation and correction/fertilization). In soil, most of the N is combined with organic matter (organic forms), but the N forms absorbed by plants are ammonium/NH4+ and nitrate/NO3− (inorganic forms). The N recommendation for agriculture crops does not observe the N available in the soil (organic or inorganic), indicating a low efficiency in nitrogen management in soil. Based on the hypothesis that the stocks of NO3− and NH4 can be used as indicative of N status in soil but with high variation according to soil factors (soil uses and management), the objective of the study was to (i) analyze the balance of nitrate and ammonium in tropical soil with different uses and management and (ii) use machine learning to explain the nitrogen dynamic in soil and the balance of nitrate and ammonium. The results showed that soil N stocks and pH promoted the formation of three clusters with the similarity between Cluster 1 (clay texture) and Cluster 2 (loam texture), represented by higher contents of nitrate as a result of high nitrification rate and lower contents of ammonium in soil. Cluster 3 (sand texture) was isolated with different N dynamics in the soil. In agricultural soils, the content of NO3− tends to be higher than the content of NH4+. There is a high nitrification rate in clay soil explained by higher organic matter and clay content that promotes soil biology. Based on the results of machine learning, for clay and loam soil, the contents of NO3 can be used as indicative of N status as a final result of nitrification rate and higher variation in soil. However, in sandy soil, NO3 can not be used as indicative of N status due to N losses by leaching.