Advances in materials characterization at the submicron and the nano-scales have progressed in the last decade. At the same time, computational capability for finite element analyses are also improving through technological developments in parallel computing. However, large computational models of nanostructured materials are currently limited by the lack of validation data. The work reported in this paper describes the formulation of a representative nanoscale model for Kevlar fibers based on failure section imaging that captures its fibril and microfibril structure. In this regard, a finite element model that captures the nanoscale structure of Kevlar fibers was developed to predict their macroscale response. Experimental derivation of geometrical parameters and physical properties of fibrils and microfibrils is challenging due to the sensitive nature of polymers. There are several microfibril parameters that reflect into effective fiber response, such as the microfibril constitutive behavior, length, diameter, shape, the interfibril shear and normal strength, and the inter-fibril normal and tangential force decay the after peak strength is achieved. This paper investigates the effect of each of the aforementioned parameters on the initial modulus, yield strength, ultimate strength, and strain rate dependence of Kevlar fibers with 10 μm average diameter. The sensitivity of the macroscale response to each microfibril parameter can be used to identify areas where experimental information can further enable the predictive capability of the computational model. A parametric study was performed to calculate the effective macroscale fiber response. Subsequently, a local gradient sensitivity method was employed to plot the sensitivity of the fiber response to each microfibril parameter.