We examine the most recent observational constraints arising from i) small-scale and large-scale Galactic dynamical properties, ii) star counts of population I and II stars at faint magnitude and iii) microlensing experiments towards the Large Magellanic Cloud and the Galactic centre. From these constraints, we determine the halo and disk stellar mass functions and stellar content down to the bottom of the main sequence, which yields the normalization of the halo/disk total stellar population, and we infer the contributions of sub-stellar objects to the mass budget of the various Galactic regions. The consistent analysis of star counts and of the overall microlensing observations in the bulge are compatible with a small contribution of brown dwarfs to the Galactic mass budget ρBD/ρ• ≤ 0.2. However the separate bulge/disk analysis based on the bulge clump giants is compatible with a substantial population of disk brown dwarfs, ∑BD/∑• ≤ 1. The most recent determination of the disk surface density, that lies within one standard deviation of all previous measurements, allows such a maximum brown dwarf contribution at the 1σ level. More statistics of microlensing events towards the Galactic center and a better determination of the velocity dispersions in the bulge should break this degeneracy of solutions. For the halo, we show that a steep mass-function in the dark halo is excluded and that low-mass stars and brown dwarfs represent a negligible fraction of the halo dark matter, and thus of the observed events towards the LMC. The nature of these events remains a puzzle and halo white dwarfs remain the least unlikely candidates. © 2004 Elsevier Science B.V., Amsterdam. All rights reserved.