Purpose To validate Monte Carlo (MC) simulations in absolute terms at the local dose deposition level for breast dosimetry applications using new experimental methodologies. Methods Three types of dose measurement devices were used for this study; thermoluminescent dosimeters (TLD-100H), metal oxide semiconductor field-effect transistor dosimeters (MOSFET, model TN-1002RD), and GafChromicTM films (model XR-QA2). First, experimental methodologies were developed and the performance of each of the three dosimeter technologies tested using a monoenergetic synchrotron-based setup. For validation under clinical conditions, a clinical digital mammography system (Mammomat INSPIRATION, Siemens Healthcare, Forchheim, Germany) and two phantoms, a semi-cylindrical homogeneous 50% glandular density phantom, and a heterogeneous anthropomorphic breast phantom, were used. MC simulations reproducing the experimental conditions were performed using code based on the Geant4 toolkit. Experimental and simulated 2D-dose maps and dose profiles for different depths in the axial-phantom plane were compared. Results The dosimeters have an energy-dependent response (it increases by ∼ 48% for XR-QA2 films and by ∼ 37% for MOSFET in the energy interval 18–28 keV), resulting in the need for calibration at each phantom depth to be investigated to obtain accurate measurements. The developed experimental methodologies allow for dose measurements with uncertainty levels of ∼ 5% (one combined standard uncertainty) at the entrance surface, up to ∼ 9% 4 cm deep inside the phantom. This increase in uncertainty is due to the decrease in dose with increasing phantom depth, which is reduced by 93% after 4 cm of the homogeneous phantom. In a heterogeneous background, dosimeters are able to detect local dose gradients of up to 46% between the boundaries of two materials. The MC-based estimates of the dose matched, to within one combined standard uncertainty, with the experimental ones for all three dosimeters technologies and for all tested depths. Conclusions The developed experimental methodologies allow for internal breast dose estimation with good accuracy. These procedures could be successfully applied to validate MC code at the local dose level and in absolute terms. The validation in a heterogeneous, anthropomorphic breast phantom is crucial to the development of a new breast model and dosimetry method.