The packing density of granular materials is in the centre of interest of many industries. For concrete industry, the packing density of aggregates is the most representative physical parameter of the granular mixtures: it combines the grain size and the morphology of grains. Compressible Packing Model (CPM) accurately predicts packing density by involving three parameters: wall effect coefficient, loosening effect coefficient and compaction index. The identification of these parameters has been made on elementary granular classes (narrow partide size distributions) which can be time consuming and difficult to achieve. In this research work, we have elaborate a strategy to determine CPM parameters for wide partide size distribution of granular mixtures. This allows reducing the number of tests to optimize packing density of granular mixtures. The results of the modelling show a good agreement with the experiment. The work will be undertaken on two typologies of grains: crushed and rolled aggregates. (C) 2019 Elsevier B.V. All rights reserved.