The paper presents a method of calculating the thermal conductivity of bio-based building composites enhanced with microencapsulated PCM to improve their thermal mass. The proposed numerical approach uses the real sample morphology for which the unsteady heat conductivity equation was solved, and a local (micro-scale) transient temperature field was obtained. The 3D sample microstructure was obtained by applying micro X-ray tomography (µCT) scanning. Then, by assuming a certain weight fraction, the microencapsulated phase change material (PCM) was randomly distributed in the space of the computational domain occupied by the binder. In this way, the artificial bio-composite with microcapsules with PCM was generated based on the real distributions of pores, hemp shives, and binder. Using artificially generated PCM-enhanced composites and based on the local transient temperature distribution, the effective thermal conductivities of the composites in the direction of heat flow were calculated by applying the volume averaging technique. In the first step, the model was tuned by selecting representative elementary volume (REV) and adjusting the predicted thermal conductivity with a measured one for bio-based composite without PCM additives. In the second step, the method was applied for predictions of thermal conductivities of composites with three amounts of microencapsulated PCM, i.e., 5, 10, and 20 % wt. The simulation results were also compared with experimental measurements, and the matching accuracy of the predicted and measured thermal conductivity values on the average level of 8.1% was achieved. © 2023 American Institute of Physics Inc.. All rights reserved.