In this study, the ability of a natural grass named Megathyrsus maximus (Guinea grass) as a sorbent for cleaning up diesel spills in water was investigated. Fourier transform infrared (FTIR) spectroscopy was used to identify the physicochemical properties of untreated and treated GG. Several parameters influencing the efficiency of oil absorbed by guinea grass were optimised using established One Factor at a Time (OFAT) and followed by Response Surface Methodology (RSM) approaches. The optimised parameters include heat treatment, time of heating, packing density, and diesel concentration, with only the significant factors proceeded to statistical optimisation through RSM. As a result of OFAT analysis, 18.83 mL of oil was absorbed at 110 °C for 15 min time of heating, with packing density of 14 g/cm3 and oil concentration of 10% (v/v). Through RSM, the predicted model was significant, confirming that packing density and oil concentration significantly influenced the efficiency of oil absorption by GG. The software predicted an oil absorption efficiency of 16.64 mL, whereas the experimental model validated the design with 22.33 mL of oil absorbed at a constant temperature and time, respectively. The RSM technique has proven better efficiency and effectiveness in absorbing oil compared to OFAT. This research advances our understanding of the utilisation of natural sorbents as a diesel pollution remediation strategy.