Three levels of three ultrasonic independent variables were optimized to obtain the maximum yield of water-soluble polysaccharides (PS) extracted from Ganoderma lucidum using response surface methodology (RSM). Box–Behnken design (BBD) was employed to evaluate the effects of ultrasonic variables on the yield of PS. The parameters that were considered for the optimization are ultrasound power (500–700 W), ultrasonic irradiation time (45–65 min) and temperature (70–90 °C). The analysis of variance suggested that the response dependent variable of yield of PS could be expressed by a quadratic polynomial model. The optimal theoretical extraction conditions were found to be an ultrasonic power of 590 W, an irradiation time of 58 min and a temperature of 81 °C. Under these conditions the predicted optimal yield was 52.33 mg. Whereas by following the optimized conditions, the yield of PS by experiments was found to be 52.28 mg which is in a very good agreement with the theoretically predicted one. These outcomes indicate the adequacy of quadratic polynomial model to represent the ultrasonic extraction variables within the ranges of investigation for a volume of 0.25 L; and any prospective scale-up may require modifications in the geometry of the extracting vessel due to the non-linear effects of power ultrasound.