In this paper, using support value-based adaptive fuzzy c-means clustering and krill herd optimization, we demonstrate how to effectively localise energy harvesting enabled underwater wireless sensor networks. Replacement or recharge of a sensor node's battery is challenging in an aquatic environment. As a result, building an energy harvester that is both efficient and dependable is essential to ensure the continued operation of an underwater wireless sensor network (UWSN). We presented a technique that is capable of harvesting energy from a variety of sources and distributing it to the sensor nodes. The proposed work gathers energy from sensor nodes with insufficient batteries and begins communicating once they have sufficient energy storage. The RSS (received signal strength) and TOA (time of arrival) of active nodes are used to determine the network's location. This is based on the characteristics of the channels used in underwater optical communication. Following that, the RSS and TOA measures' support values are determined. Then, using support value-based adaptive fuzzy c-means clustering, support kernel matrices are created. The proposed support kernel matrices significantly reduce path error during data transmission. To increase sensor node localisation, the obtained support kernel matrices are further improved using a krill herd optimization approach. The proposed method outperforms existing techniques in the laboratory.