Different ammonium nitrate controlled-release (CR) systems based on ethylcellulose have been investigated to reduce environmental pollution derived from nitrogen-fertilizer use. Coated ammonium nitrate granules were produced in Wurster-type fluidized-bed equipment using two different amounts of ethylcellulose. The highest one was modified by the addition of two plasticizers, dibutyl sebacate, and dibutyl phthalate. Having researched the encapsulation efficiency and the homogeneity of the coated granules, we carried out the kinetic-release experiments in water and soil. The release rate of the active ingredient was related to the thickness of the coating film, granule size, and type of plasticizer used. Using an empirical equation, the time taken for 50% of the active ingredient to be released into water and soil (T-50) was calculated. From the analysis of the T-50 values, we can deduce that the release rate of ammonium nitrate can be controlled, mainly changing the thickness of the coating film and using plasticizer as well. In water experiments, T-50 values for granules prepared without plasticizers ranged between 7.47 h for 1 mm < d < 2 mm granules coated with 10% of ethylcellulose and 24.06 h for 2 mm < d < 3 mm granules coated with 20% of ethylcellulose. For those prepared with plasticizers, T-50 ranged between 22.80 h for 1 mm < d < 2 mm granules containing dibutyl sebacate and 35.74 h for 1 mm < d < 2 mm granules containing dibutyl phthalate. However, in soil experiments T-50 values ranged between 10.24 h for 1 mm < d < 2 mm granules coated with 10% of ethylcellulose and 38.80 h for 1 mm < d < 2 mm granules containing dibutyl sebacate. Finally, a linear regression of the T-50 values was obtained by the results of the study carried out in water and soil. This allows us to predict the behavior of the formulations in soil. This could be useful in the design of systems which control the nitrogen release.