Extracts from Hibiscus sabdariffa L. (roselle) have been used traditionally as a food, in herbal medicine, in hot and cold beverages, as flavouring or coloring agent in the food industry. In vitro and in vivo studies and trials provide evidence, but roselle is poorly characterised phytochemically due to the extraction processes. The optimization of the extraction of phenolic compounds and their antioxidant activities is still a hot topic. In this study, the effect of solute/solvent ratio (33, 40 and 50 mg/mL), extraction temperature (40, 50 and 60 °C) and extraction time (30, 60 and 90 min) was evaluated through the content of phenolic compounds and antioxidant activity. A response surface methodology through a Box–Behnken design was applied and model fit, regression equations, analysis of variance and 3D response curve were developed. The results showed that TPC, TFC, DPPH and FRAP were significantly influenced by temperature, extraction time and solvent/solute ratio. Thus, TPC, TFC, DPPH and FRAP varied from 5.25 to 10.58 g GAE/100 g DW; 0.28 to 0.81 g QE/100 g DW; 0.24 to 0.70 mg/mL; 2.4 to 6.55 g AAE/100 g DW respectively. The optimal experimental condition (41.81 mg/mL; 52.35 °C and 57.77 min) showed a significant positive effect compared to conventional methods. The experimental values at this extraction condition show that this optimization model is technologically, financially and energetically viable as it requires a reasonable concentration, time and temperature.