Ammonium nitrate fuel oil is an explosive mixture found in most antipersonnel landmines (APL) buried throughout the Colombian territory. During more than 50 years of internal conflict, the Colombian government has found that trained dogs are the most effective method to detect APL. However, the olfactive signature in ANFO is unknown and also if there are differences in detection related to the explosive manufacturing origin. Therefore, this work begins with the analytical validation of the method used to determine ammonia, in its derivatized form as carbamate, released by home‐made ANFO using HS‐SPME‐GC‐FID. Once validated, the method was used to identify ammonia and other organic volatile compounds present in ANFO, under laboratory and simulated field conditions. The validation process includes the evaluation of the optimum conditions for the derivation and extraction of butylcarbamate, the determination of the working ranges with linear response in FID, the limits of detection and quantification, the sensitivity, and the precision. The results of the validation established linearity and sensitivity in a concentration between 20 and 120 mg/L, as well as low limits of detection and quantification of 6.4 and 21.4 mg/L, respectively. Also, an intermediate precision of 11% for butylcarbamate with a repeatability of 8%. The validated method showed in real samples of home‐made ANFO besides ammonia, the presence of low molecular methylamines, and also exhibited differences in volatile compositions according to the origin. The objective of this work is to offer a reliable analytical methodology for the extraction and analysis of volatile compounds from ANFO.