To describe a low-field nuclear magnetic resonance (NMR) method for quantifying pore-fracture fractal dimensions and their influence on effective porosity and permeability, we performed modeling comparisons between fractal analysis and pore-fracture physical properties in low-rank coals. The adsorption space fractal (D-NMRA), seepage space fractal (D-NMRS) and moveable fluid space fractal (D-NMRM) were calculated to be 1.62-1.91, 2.77-2.98 and 1.56-2.75, respectively. The D-NMRA generally increased with increasing Langmuir volume (V-L, 9.54-31.06 m(3)/t), Langmuir pressure (P-L, 0.58-8.13 MPa), the Brunauer-Emmett-Teller (BET) surface area and its fractal dimension. Higher D-NMRA indicated the significant coalbed methane (CBM) adsorption capability. Both the D-NMRS and D-NMRM decreased with increasing areas of T-2 > 2.5 ms distribution (S-T and S-CT) and sorting coefficient. These phenomena showed that the NMR fractal method could reflect the coal pore-fracture heterogeneity and had significant influence on seepage space content. The correlations of moveable fluid porosity and permeability with D-NMRM can be found by performing the models of y = ax + b (a < 0), so coals with high D-NMRM occur to have low flow capability. Furthermore, the pore-fracture porosity and permeability have positive correlations with ST and SCT, which result from the connection between pores and fractures. These results also show that fractal analysis calculated with T-2 can be developed to appraise the physical properties of low-rank coals and supply some reference for a relatively full identification of porous media. We advise that low-field NMR can be employed as a lossless analytic method to quantify moveable fluid space fractal theory. (C) 2016 Elsevier Ltd. All rights reserved.