In Fifth Generation (5G) network, small cells have been incorporated to meet the growing demands for mobile traffic and ubiquitous service. It also enables Internet of Things (IoT) applications that will be positioned all over with mobile broadband technology. As modern IoT applications have active and varied necessities, several systems arrange mixed sorts of network sources consequently. Prevailing studies recommend methods for sustaining the distribution of network segments to diverse applications proficiently.Balancing the loads amongst the cells turns out to be one of the significant problems in cellular network. In this paper, a load balancing technique based on network segmentation and adaptive sleep scheduling for 5G-IoT networks is proposed. In this technique, for each network segment,sub-segments are formed and grouped to process IoT application shaving different Quality of Service (QoS)requirements. The improved DBSCAN algorithm is used for sub-segmentation and grouping. based on a combined rank of QoS factors. In the next phase, each small cell base station (SBS) executes adaptive dynamic sleep scheduling based on its load level. In the load balancing policy of SBS, when the average load of any SBS increases beyond the load of the Macrocell, the overloaded traffic can be moved to the Macrocell. Simulation results are validated against the analytical results and hence prove that the proposed LBNSASS technique achieves higher success probability and power efficiency with reduced energy consumption and packet drops of SBSs when compared to existing technique.