Laser Doppler flowmetry (LDF) can be used for assessing the microcirculatory perfusion. However, conventional LDF (cLDF) gives only a relative perfusion estimate for an unknown measurement volume, with no information about the blood flow speed distribution. To overcome these limitations, a model-based analysis method for quantitative LDF (qLDF) is proposed. The method uses inverse Monte Carlo technique with an adaptive three-layer skin model. By analyzing the optimal model where measured and simulated LDF spectra detected at two different source-detector separations match, the absolute microcirculatory perfusion for a specified speed region in a predefined volume is determined. qLDF displayed errors<12% when evaluated using simulations of physiologically relevant variations in the layer structure, in the optical properties of static tissue, and in blood absorption. Inhomogeneous models containing small blood vessels, hair, and sweat glands displayed errors<5%. Evaluation models containing single larger blood vessels displayed significant errors but could be dismissed by residual analysis. In vivo measurements using local heat provocation displayed a higher perfusion increase with qLDF than cLDF, due to nonlinear effects in the latter. The qLDF showed that the perfusion increase occurred due to an increased amount of red blood cells with a speed>1 mm∕s.