To analyze and predict complex crack propagation behavior in materials, numerical simulation is an effective methodology. Among the available simulation methods, the phase-field fracture (PFF) model has attracted particular attention due to its ability to describe crack initiation, propagation direction, coalescence, and branching. The predictive accuracy of PFF simulations depends on the physical properties and parameters included in the model and the constitutive laws of the material. Recently, Bayesian data assimilation, which integrates experimental data into numerical simulation data, has gained attention as a robust method to inversely estimate the numerical model parameters and improve the predictive accuracy of simulation. By applying the Bayesian data assimilation to PFF simulations, it becomes possible to improve the precision of crack propagation prediction by integrating experimental data into PFF simulation results. This study aims to develop a data assimilation methodology for improving the accuracy of PFF simulation of fatigue crack propagation by integrating the simulation results with non-contact strain measurements obtained by the digital image correlation method. In this presentation, the results of numerical experiments conducted to validate the data assimilation methodology constructed in this study are presented.