Abstarct: Some basic properties for negatively superadditive dependent NSD random variables are presented‎, ‎such as the Rosenthal-type inequality and the Kolmogorov-type exponential inequality‎. ‎Using these properties‎, ‎we further study the complete convergence for weighted sums of NSD random variables‎, ‎which generalizes and improves some corresponding ones for independent random variables and negatively associated random variables‎. ‎Some sufficient conditions to prove the complete convergence for weighted sums of NSD random variables are provided‎. ‎As an application‎, ‎the complete consistency of LS estimators in the EV regression model with NSD errors is investigated under mild conditions‎, ‎which generalizes and improves the corresponding one for negatively associated random variables‎. ‎Finally‎, ‎we will evaluate theoretical results with the simulation study‎. ‎