Today, Most smart mobile devices are facilitated with advanced processing hardware and short-range data communication systems by which they are practically capable to provide effective execution services to the neighbor mobile device client request and/or receive services on a need basis within the local area network. Therefore, to relish the powerful capability of these smart mobile devices in the private campus network, we propose an intelligent composite offload decision algorithm (ICODA) framework that attempts to connect several smart mobile devices in wireless local area network and make them apply intelligence before servicing each other request preferably without the internet. The significance of the proposed framework is that it has a mechanism to make a data offloading decision using an optimal decision tree classifier model and also a mechanism to avoid data offloading operation using the data cache neural networks model. The experimental results obtained are obvious to show the minimal client system battery utilization and hence an optimized work time for a smart mobile client device that participates in the ICODA framework.