<p>Seepage-induced disasters pose a significant problem in loess regions. The permeability coefficient of loess serves as one of the key parameters for analyzing such problems. The seepage process in unsaturated loess can trigger internal pore collapse, which exerts a significant influence on the permeability coefficient. In this paper, a new analytical model for the permeability coefficient of unsaturated loess considering internal pore collapse is proposed, and the influence of internal pore collapse on the permeability coefficient of loess is systematically analyzed. First, the effect of pore collapse on soil–water characteristic curve (SWCC) is investigated. Two parameters, namely the compression factor and the translation factor, are introduced to describe the influence of pore collapse on the pore volume function, and a modified SWCC model incorporating pore collapse effects is established. Then, an analytical model for permeability coefficient of loess is developed based on the permeability method proposed by Brooks and Corey. The proposed model in this paper contains nine parameters, all of which can be determined through laboratory test fitting. Using an initial reference void ratio of e = 0.85, the model parameters are calibrated, and the permeability coefficient of unsaturated loess at different pore states are predicted. Finally, the sensitivity of the model parameters to the prediction results is discussed.</p>