Despite using encryption protocols such as HTTPS, web page fingerprinting poses significant privacy risks, even when traffic analysis is used to identify specific web pages visited by users. Adversaries can exploit packet-level characteristics like packet length to gather information about user behaviour and preferences without decrypting traffic. This paper uses encrypted traffic packet lengths to distinguish webpages based on privacy-driven fingerprinting – FineWP class webpages based on packet length sequences in a bidirectional client-server interaction. Our results demonstrate that FineWP outperforms traditional and deep learning-based methods regarding runtime and accuracy. Based on our experimental results, FineWP demonstrates robust and privacy-protected fingerprinting capabilities for fine-grained webpage identification, effectively managing large-scale datasets consisting of numerous webpages and substantial background traffic. We propose an innovative webpage fingerprinting method that exclusively utilizes encrypted packet length information, achieving an impressive accuracy of 94.3% while rigorously preserving user privacy. Additionally, our lightweight and efficient technique exhibits strong resistance against sophisticated traffic analysis attacks, significantly outperforming existing deep learning-based fingerprinting approaches by approximately 11.2% in terms of accuracy, computational efficiency, and resilience under realistic network conditions. These findings highlight the potential of FineWP for secure, scalable, and practical webpage fingerprinting applications.