We present a simple, largely empirical but physically motivated model, which is designed to interpret consistently multi-wavelength observations from large samples of galaxies in terms of physical parameters, such as star formation rate, stellar mass and dust content. Our model is both simple and versatile enough to allow the derivation of statistical constraints on the star formation histories and dust contents of large samples of galaxies using a wide range of ultraviolet, optical and infrared observations. We illustrate this by deriving median-likelihood estimates of a set of physical parameters describing the stellar and dust contents of local star-forming galaxies from the Spitzer Infrared Nearby Galaxy Sample (SINGS) and from a newly-matched sample of SDSS galaxies observed with GALEX, 2MASS, and IRAS. The model reproduces well the observed spectral energy distributions of these galaxies across the entire wavelength range from the far-ultraviolet to the far-infrared. We find important correlations between the physical parameters of galaxies which are useful to investigate the star formation activity and dust properties of galaxies. Our model can be straightforwardly applied to interpret observed ultraviolet-to-infrared spectral energy distributions (SEDs) from any galaxy sample.