This study extends a traditional problem on locating distribution centers (DC) under uncertain demands, and proposes a stochastic programming model on locating DCs, determining their suitable scales, as well as adjusting DCs so as to adapt to the dynamically changing demands at the sites of retailers. Some lower and upper bounds are also designed to analyze the values of the perfect information, stochastic solution, and the DC adjustments. Numerical experiments are conducted to validate the effectiveness of the proposed model.