In response to the burgeoning interest in Vehicle-to-Everything (V2X) communications and its pivotal role in enhancing road transportation efficiency and safety, this paper presents a novel scheme for the formation of vehicle platoons, addressing the critical need for realistic implementation strategies. Recognizing the limitations of existing research that often relies on idealized assumptions, this work proposes a platoon formation method utilizing the real-world dataset. The method combines the robustness of Random Forest algorithms for feature selection with the adaptability of Fuzzy K-Means (FKM) clustering to effectively group vehicles with similar characteristics into platoons. This approach ensures the consistency of speed and acceleration within platoons, a crucial aspect for maintaining platoon integrity and stability. Validated through the Kaggle dataset, the scheme not only showcases the feasibility of forming efficient vehicle platoons in realistic scenarios but also introduces a velocity and acceleration-based Cooperative Adaptive Cruise Control (CACC) strategy to prove the formatted platoon stability.