<p>Freshwater scarcity is still a global problem, particularly in desert and coastal regions where conventional desalination methods are costly, energy-intensive, and harmful to the environment. The Tubular Solar Still (TSS) desalination technique using solar energy has been regarded as a renewable technique with lower costs, though the higher initial investment costs are a disadvantage; the disadvantage is that the rate of freshwater production is lower.To address these issues, this research proposes a new hybrid intelligent approach called Chaotic Banyan Rhizome-Wavelet Kernel Extreme Learning Machine (CBR-WKELM) for performance improvement and precise prediction of TSS dynamics.On the experimental side, two systems were designed and tested as a Conventional Tubular Solar Still (CTSS) and a Modified Tubular Solar Still (MTSS). The MTSS was further equipped with a Photovoltaic (PV)-powered electrical heater at the bottom of the absorber plate. The key operational parameters, such as solar radiation, ambient temperature, water temperature, absorber plate temperature, cover plate temperature, and distillate output, were also recorded.The experimental outcome showed that the MTSS achieved a distillate yield of 33.3% higher, thermal efficiency of 31.0% higher, and exergy efficiency of 29.3% higher than the CTSS. In the modeling process, the Chaotic Hiking Optimization Algorithm (CHOA) was employed for optimal feature selection, while the Wavelet Kernel Extreme Learning Machine (WKELM), fine-tuned by the Banyan Tree with Binary Plant Rhizome (BT-BPR) algorithm, achieved outstanding predictive accuracy with Root Mean Square Error of 0.28, Mean Absolute Error of 0.23, and Coefficient of Determination of 0.98, outperforming the Artificial Neural Network (ANN) and Particle Swarm Optimization-based ANN (PSO-ANN).</p>