Using monthly streamflow data in Dongjiang Basin, South China, hydrological drought events were truncated on the basis of the run theory. The multivariate joint distributions of droughts were determined by using the function of Meta-Gaussian Copula. An improved method for multivariate design, the most-likely weight function, was applied to select design values of drought properties for the given Kendall joint return periods. Besides, impacts of water reservoirs in the joint design values of drought properties were quantified. Results demonstrate that the Weibull distribution fit the marginals of drought properties well. The dependences of drought properties were highly positive and the Meta-Gaussian Copula effectively modelled their bivariate and trivariate joint distributions. Despite of the bivariate or trivariate design, a set of design values of drought properties almost occurred in the identical cumulative frequencies with slight difference. The remarkable decrease in drought properties of the Dongjiang Basin was attributed to the regulation of water reservoirs. The exceeding joint return periods significantly increase for a set of drought properties under the impacts of water reservoirs. The shorter the joint return period was, the larger the change of the design values was. However, the duration, severity and peak of drought for the joint return period by 10 years still accounted for 3.89-4.04 months, 7.20-7.97 and 2.99-3.12 hundred-million m³, respectively, provided that minimum instream flow was satisfied for water demand in the Dongjiang Basin.