We present a new approach for analysis of the dust emission spectra of the central infrared condensations in the dense cores of molecular clouds AFGL 2591, B335, L1551, Mon R2 and Sgr B2. We find that the far-infrared spectra of these objects can be reasonably well characterized by the functional form F(nu) Anu(beta)e-Bnu1/2, where A, beta, and B are positive parameters. Assuming a power-law wavelength dependence of the dust emissivity and that the dust emission is optically thin within the frequency range considered, we derive analytically the dust temperature distributions and total dust masses for these objects using an inversion method. The dust temperature distributions obtained in all cases indicate the existence of a low-temperature limit and a characteristic average temperature. The amount of dust at temperatures much higher than the characteristic average temperature decreases drastically, and most of the dust is within a few tens of degrees above the low ''cutoff'' temperature. The low-temperature cutoff and the characteristic average dust temperature are naturally sensitive to the power-law exponent for the emissivity law, indicating that our inadequate understanding about the emissivity at long wavelengths is the major source of uncertainty in the framework of this inversion technique. We compare our results with previous investigations using single-temperature dust models.