Precise Web page identification has always been a research hotspot in the areas of network management and security. However, previous works generally focused on statistical or probabilistic approaches and could not exactly calculate the length of encrypted data under different conditions, which makes them hardly cover all the cases. In this poster, we propose an exact fingerprint derivation method for encrypted Web-browsing traffic and thereby implementing a prototype system for precise page identification (PPI). Our experiments show that PPI not only can be employed for early page identification at individual-flow level but also can achieve very high accuracy at aggregate-traffic level.