The formulation of alkali activated-binders from numerous waste and industrial by-products is gaining popularity in concrete technology to reduce waste dumped into landfills. This study investigated the effects of incorporating varying proportions (25%–40%) of waste glass powder (WGP) into ground granulated blast furnace slag (GGBS) with constant proportion (10%) of metakaolin (MK) on the workability, mechanical and microstructural characterizations of ternary blended geopolymer binder. The fraction of alkaline activator solution/binder composition, sodium silicate solution to sodium hydroxide solution, and the concentration of sodium hydroxide solution used were 0.55, 2.5, and 2M–12M, respectively. The experimental results demonstrated that the optimum replacement percentage of WGP in ternary-based geopolymer concrete was observed to be 35%; subsequent increment in WGP substitution levels resulted in the progressive decrease in mechanical properties. Furthermore, the mineralogical (XRD) and microstructural (SEM-EDS) characterizations of the geopolymer samples revealed the formation of calcium silicate hydrate (C–S–H), calcium aluminate silicate hydrate (C-A-S-H), and sodium aluminate silicate hydrate (N-A-S-H) gels as the end reaction products. Consequently, this study developed an Artificial Neural Network (ANN) framework to assess the workability and mechanical properties of ternary blended geopolymer binder employing WGP replacement levels and varying concentrations of sodium hydroxide solution as input parameters. The findings revealed that the ANN framework could be an efficient approach for predicting the workability and mechanical properties of the ternary-based geopolymer matrix.