This study aims to develop reliable and practical models for estimating the compressive strength and CO2 emission of geopolymer concrete (GPC) using the Gene-expression programming (GEP) technique. Approximately 274 data points on the strength and CO2 emission of GPC using fly ash (FA) and ground granulated blast furnace slag (GGBFS) were used to generate two prediction models considering the effects of 19 input variables. The research results show that the proposed GEP-based models perform well with high correlation coefficients (R > 0.95) and reasonable errors. The parametric study demonstrates that increasing the GGBFS content, the Na2O (dry) content, and the molarity of NaOH solution, and decreasing the FA content, the FA/GGBFS ratio, and the added water or total water/binder ratio could significantly enhance the strength of GPC. Besides, the proposed GEP-based models were used to optimise the mix proportions of GPC concrete with minimisation of CO2 footprint. The range and reasonable values of the FA and GGBFS contents, the alkaline activators, the aggregates, and other parameters were suggested based on the in-depth analyses. This significant reference data source could promote the use of GPC widely in the construction industry as green construction materials for sustainable development.