Various wall materials, including maltodextrin (MD),beta-cyclodextrin (beta-CD), whey protein isolate (WPI) and gum Arabic (Gum-A) were combined together as the wall materials for encapsulation of blueberry anthocyanin extracts through freeze drying. Simplex lattice mixture design was used to make the experimental design. Artificial neural network (ANN) combined with genetic algorithm (GA) was successfully applied to model the influences of formulation composition on encapsulation productivity (EP) and encapsulation efficiency (EE), as well as obtain optimum formulations. Four optimum formulations were provided by the ANN-GA approach. In all the optimum formulations, WPI had the highest content. Using the optimum formulations, EP values were higher than 96% and EE values exceeded 82%. On the other hand, the properties of resulting anthocyanin powders were analyzed from different aspects. Although there were some differences in bulk density, particle size and glass transition temperature among encapsulated powders using different formulations, all the samples exhibited similar moisture content, water activity, color property and crystallinity. More importantly, encapsulation using resulting optimum formulations was effective to protect blueberry anthocyanins against degradation during heating. (C) 2017 Elsevier B.V. All rights reserved.