This paper deals with the lifetime problem in the Internet of Things. We first propose an efficient cluster-based scheme named "Cuckoo-search Clustering with Two-hop Routing Tree (CC-TRT)" to develop a two-hop load-balanced data aggregation routing tree in the network. CC-TRT uses a modified energyaware cuckoo-search algorithm to fairly select the best cluster head (CH) for each cluster. The applied cuckoo-search algorithm makes the CH role to rotate between different sensors round by round. Subsequently, we extend the CC-TRT scheme to present two methods for constructing multi-hop data aggregation routing trees, named "Cuckoo-search Clustering with Multi-Hop Routing Tree (CC-MRT)" and "Cuckoo-search Clustering with Weighted Multi-hop Routing Tree (CC-WMRT)." Both CC-MRT and CC-WMRT rely on a two-level structure; they not only use an energy-aware cuckoo-search algorithm to fairly select the best CHs but also adopt a load-balanced high-level routing tree to route the aggregated data of CHs to the sink node. However, CC-WMRT slightly has a better performance thanks to its low-level routing strategy. As an advantage, the proposed schemes balance the energy consumption among different sensors. Numerical results show the efficiency of the CC-TRT, CC-MRT, and CC-WMRT algorithms in terms of the number of transmissions, remaining energy, energy consumption variance, and network lifetime.