AbstractComplex failure modes of carbon fiber reinforced polymer (CFRP) composites create significant challenges for CFRP components' progressive damage analysis (PDA). Though important in PDA, the phenomenological strain‐softening model (SSM) for characterizing the degradation behavior in the fiber‐parallel direction (dominated by fiber breakage) of CFRP is difficult to be constructed by current approaches due to the poor toughness of carbon fibers. The work aims to obtain a refined SSM of plain woven CFRP composites through experimental methods and apply it to the PDA of corresponding components. A novel phenomenological cyclic‐loading‐based strain‐softening model (CLSSM) was proposed based on the theory of equivalent dissipation energy during static and cyclic loading. Firstly, the release behavior of dissipation energy during the cyclic loading process was researched. Secondly, a newly designed search algorithm taking degraded mechanical properties during cyclic loading as input and taking the static strain‐softening curve as output was utilized to build up CLSSM. Thirdly, a three‐point bending test was conducted and a corresponding numerical model was constructed for comparison to verify the proposed model. The entire framework is highly programmatic, and results show that CLSSM is fast for characterizing the strain‐softening behavior, and effectively simulates the progressive failure process of CFRP components.