A wireless sensor network (WSN) consists of independent sensors spread across different locations, tasked with monitoring environmental or physical parameters such as sound, pressure, temperature, and more. These sensors communicate wirelessly and often collaborate to perform sensing tasks efficiently. Nevertheless, these networks are susceptible to numerous security threats. A novel approach called Security-Enabled Blockchain with Improved Rat Swarm Algorithm (SEBIRA) has been proposed to address these challenges. This method enhances network security while minimizing energy consumption in WSNs. The SEBIRA method integrates three phases to achieve superior network performance and security. Initially, SEBIRA employed a quad-tree structure for network organization, simplifying management and reducing complexity. Secondly, Improved Rat Swarm Optimization (IRSOA) is used to dynamically select optimal cluster heads (CH), enhancing communication efficiency and reducing energy usage. Thirdly, the Convolutional Neural Network (CNN) algorithm is employed to implement duty cycling, thereby conserving energy and increasing the network’s lifespan. Subsequently, secure classification is performed using empirical mode decomposition with a bi-directional long short-term memory (EMD-BILSTM) network. Simulations conducted using NS-3.26 validate the superiority of this approach over existing methods, demonstrating high detection accuracy, maximum throughput, minimal latency, and reduced energy depletion. The SEBIRA technique extends network lifetime significantly compared to the EPFMR, SECDL, and GAPSOH methods by 8.7%, 6.8%, and 9.97%, respectively, underscoring its effectiveness in enhancing WSN performance and security.