Proven to be a viable and cost effective solution for ubiquitous internet access, wireless access network have proliferated at homes, offices, and cafes all over the globe. Being placed at the very edge of the computer network, with increasingly powerful computational ability, wireless access points are more and more suitable for deploying configurable and personalized services for mobile clients. In this paper we propose an intelligent wireless service framework over enhanced wireless access points (refered to as Smrart Wireless Access Point) that prefetch and aggregate web contents on the wired side based on user profiles and delivery it to the user in a wireless network environment. We modify the source code of browser to enhance its data collection capabilities and develop some practical smartphone applications on the mobile client-side. Additionally, we perform some caching stratagy and algorithms when a user have created considerable usage data or rated enough information. Measuring the download latency shows that DNS caching and content caching yields significant improvements in page load time. Prefetching mechanism, recommend system and some rich featured wireless local applications all benefit users much.