A three-level Box-Behnken design of response surface methodology was employed to optimize the formulation of bukayo or coconut toffee based on overall acceptability sensory attribute. The independent variables were different levels of young coconut meat (250 -500 g), sinakob (250 -500 g), and coconut juice (250 -500 g) with fifteen different runs of three similar center points. The analysis of variance showed that quadratically, coconut juice and sinakob showed significantly higher effect (p<0.05) on the overall acceptability of bukayo whereas the "Lack of Fit" was non-significant relative to the pure error. For optimization, the goal was to maximize the overall acceptability response. The optimum formulation conditions of 430 g young coconut meat, 400 g sinakob, and 340 g coconut juice yielded an experimental overall acceptability score of 8.04 which is close to the predicted score of 7.11 with composite desirability of 0.87 on a nine-point hedonic scale. The regression model equation for overall acceptability could predict response yield until at least 84.65% of accuracy. Box-Behnken design of RSM performed well in the optimization of the bukayo formulation with a very high degree of overall acceptability. The strong and positive correlations enable the application of colorimetry as an indication of the overall acceptability and sensory color of the bukayo product.