Continuous fiber reinforced Polymer Matrix Composites (PMC) are replacing traditional steel for different load bearing applications due to their high specific strength and stiffness. Designing these structures requires the knowledge of strength and stiffness properties of FRP composites. Micromechanical modeling through finite element analysis of representative volume element has been widely used by the researchers for prediction of elastic and strength properties, and also for studying different damage mechanisms like matrix cracking, fiber breakage and interface failure. In literature, both random and periodic fiber distribution models exist. While effective properties can be predicted by periodic models, random models are required to study damage initiation and progression. UDFRP composites also exhibit statistical variations in the material properties. These variations in the material properties arise due to the intrinsic behaviour of composites like inhomogeneity, anisotropy and variability induced during fabrication. It is essential to acknowledge these variations and use a statistics-based strength property for designing composite structures. There is a lack of literature on prediction of statistics of strength distribution through micromechanical modeling. In the present work, the effect of non-uniformity in inter fiber distance on statistics of strength distribution of UDFRP composite is studied through computational micromechanics. Micromechanics based realizations of the UDFRP composite are developed with a quantified measure of non-uniformity (MoN) in fiber distribution through an algorithm. These realizations are subjected to transverse load. A modified Drucker Prager failure criterion are implemented for matrix failure. Statistical distribution of failure initiation strength is obtained for each fiber distribution. Variation of mean value, standard deviation and B-basis of failure initiation strength with the measure of non-uniformity in fiber distribution is studied for fiber volume fractions 0.5 and 0.6.