Abstract: Soft materials such as hydrogels and rubbers have attracted increasing attention in both fundamental research and industrial applications due to their exceptional stretchability, reversible deformability, and outstanding energy dissipation capability. However, fracture remains one of the most prevalent and destructive failure modes in such materials, and accurately predicting crack initiation, propagation, and ultimate rupture is crucial for ensuring the safety, reliability, and longevity of soft-structure systems. In this study, a finite-strain phase-field model for thin-film fracture is developed to elucidate the underlying mechanisms of crack interaction and to investigate how material parameters influence crack evolution paths. The proposed model, specifically formulated for hyperelastic thin films such as polydimethylsiloxane (PDMS), provides a comprehensive framework that incorporates the derivation of governing equations, numerical solution strategies, and an adaptive mesh refinement scheme, thereby achieving an optimal balance between computational accuracy and efficiency. The predictive capability of the model is verified through direct comparison between numerical simulations and experimental tensile tests on single-edge-notched specimens, demonstrating strong consistency in both crack propagation morphology and load–displacement responses. Furthermore, for double-edge asymmetric configurations, the model is employed to systematically analyze the influence of strain softening and hardening behavior, initial crack length, and vertical spacing on crack deflection. The results reveal that stress-field coupling between interacting cracks significantly modifies the local stress distribution near crack tips, inducing pronounced path deviation, while the initial crack length has only a minor effect on trajectory evolution. Overall, the proposed model effectively captures multi-crack evolution behavior in thin films, offering a robust theoretical and numerical foundation for fracture prediction in flexible polymers and other nonlinear materials.