The advancement in technology has created smart cities, smart gadgets, and other applications where the data get collected form the wireless sensor networks (WSN) for the analysis of requirements of the users. The sensor nodes are user for collecting the soil, water, and temperature data of the agriculture filed for the automatic analysis of agriculture land condition. The existing methods tried to automate the analysis of the agricultural filed to analyze the crop productivity but failed to reach the expected outcome due to energy drain in the sensor nodes. To solve the existing problem the Roulette wheel-based multiverse optimization algorithm with opposition-based learning (RWMVO with OBL) technique is used for the selection of fittest sensor nodes from the available resources that increased the life span of the sensors for multidata collection task for smart agriculture. The OBL algorithm was incorporated for advancing the learning ability of RWMVO. The performance of the developed RWMVO with OBL has obtained higher throughput of 5.68 Kbps, lower energy utilization of 9.05 mW, and less time delay of 795 ms for 500 rounds compared to the existing energy aware software-defined network (EASDN) model.