In this presentation, strength estimation of a unidirectional fiber-reinforced composite material by peridynamics, considering random field modeling of microstructure and microscopic in-situ measurement, is discussed for the quantitative apparent strength estimation of the material. The randomness and uncertainty in the microstructure of composite materials, particularly the random variation in the inclusion locations, have a significant influence on the microscopic stress distribution and lead to dispersion in their apparent strength. Conventional methods like the finite element method are used for stress analysis in composites, but there are some difficulties in precisely modeling of complex microstructures and damage propagation, especially when considering multiple cracks caused by the interfaces between many reinforcements and base materials. For this problem, alternative approaches like XFEM and cohesive-zone element analysis have been attempted, but more effective approaches are desired. For this background, this study attempts to employ peridynamics. This approach is particularly adept at handling multiple inclusions and discontinuities like cracks in the target media, rather than conventional analytical methods. In particular, the study incorporates random field modeling utilizing actual composite specimen images to generate realistic fiber arrangement, and the peridynamics approach with the microscopic in-situ observation of the displacement field and failure condition in the microscale, previously proposed by the authors. In the presentation, details of the problem setting, peridynamics, and the random field modeling methodology will be introduced. Further, the results of microscopic fracture analysis of the target material using the presented peridynamics-based approach with experimental microscopic in-situ observations and measurements will be discussed. In conclusion, the validity and effectiveness of the presented approach for a quantitative evaluation of the materials will be discussed.