A key limitation in existing research, namely the neglect of the capability of connected and automated vehicle (CAV) was addressed in order to receive feedback from multiple preceding vehicles beyond direct physical constraints. An improved car-following model, termed I-CACC, was proposed by incorporating multi-predecessor information feedback in order to evaluate the impacts of varying CAV penetration rates and vehicle platoon configurations on mixed traffic flow stability and fundamental diagram property. The spacing differentials between multiple preceding vehicles and the subject vehicle were employed as a criterion to quantify influence weights. A platoon configuration model was developed based on platoon intensity and Markov chain theory. The platoon stability and fundamental diagram property of mixed traffic flow were simulated and analyzed. Results showed that the stability of the mixed traffic platoon was maximized when the CAV penetration rate reached 0.7 and the platoon intensity was 0.5. Stability deteriorated beyond these thresholds. The capacity of a pure CAV flow with multi-predecessor feedback increased by 3.09 times compared with that of a pure human-driven vehicle (HV) flow, whereas the improvement was limited to 2.36 times when adopting the conventional CACC model in pure CAV flow.