Introduction. New ecological agricultural practices contribute to improved pest management and root development of pineapple, but tools to quantify the root development in cultural profiles in situ are lacking. The spatial distribution of the root length density (RLD) is a key factor for the absorption of water and nutrients. A robust model was sought to predict the RLD from the number of intersections of roots per unit area (RID) in a soil profile (trench method). Materials and methods. The procedure was based on the extraction of cubes of 1 dm3 of undisturbed soil on profiles under pineapple plants in triplicate, in which the RID was counted on three perpendicular faces of the cubes and the RLD was measured inside the cubes. Results and discussion. A model predicting the RLD (RLDc) from counting of the RID was developed and successfully tested: RLDc = RID.CO.CE, where CO = 2.65 is an orientation coefficient, and CE = 1.69 is an empirical coefficient. These two coefficients are fixed. Conclusion. The model allows an estimate of pineapple RLD and its spatial variability from simple counts of roots on a soil profile. A practical example of this model is given, characterizing and comparing in situ root profiles of pineapple plants.