Slag viscosity is key factor of liquid steel refining and the smooth slag tapping of entrained flow coal gasifiers, and the accurate estimation of slag viscosity can provide a guidance for the selection of industrial raw material and the control of industrial process. In this study, a structure-related viscosity model about the multicomponent slags of the CaO-SiO2-Al2O3-FexO system is developed based on the flow mechanism of silicate melt to improve slag viscosity behavior. The present estimation model improved the expression of activation energy equation on the basis of the model proposed by Nakamoto and expanded the application from the ternary CaO-SiO2-Al2O3 system to the quaternary CaO-SiO2-Al2O3-FexO system. Then, the structural roles of all the components were clarified, and the types of oxygen bonds were defined by Raman spectroscopy. The fractions of various oxygen bonds were obtained, and the database of oxygen bonds in any composition within the range of experimental slags was generated by utilizing the Lagrange interpolation method. The result showed that the present model had an advantage in estimating slag viscosity within the compositional range of this experiment slag, and the mean deviations between the viscosities calculated by the present model and measured or published viscosities were within 20%. The model will offer a reference for the estimation and adjustment of slag behavior in the metallurgical slag or coal slag, especially for the iron-rich system.