<p>The growing need for sustainable construction materials and the environmental impact of cement-intensive products necessitate finding sustainable ways to provide strength, durability, and cost-effectiveness with products enabling large-scale use of industrial by-products. However, the performance variability of waste-based binders and the lack of reliable predictive instruments hinder their extensive use. This research examines the feasibility of fly ash (FA) and ground granulated blast furnace slag (GGBS) as clean binder systems through experimental and predictive modelling. Experimental tests were conducted on unconfined compressive strength (UCS), water absorption, drying shrinkage and durability of cyclic wetting–drying cycles in FA-GGBS-lime and FA-cement mixes. Results show enhanced performance of optimised ternary blends, FA + 20% + 20% GGBS, achieving UCS values over 108.8 kg/cm², exceeding traditional brick strength requirements. To address performance predictability, Artificial Neural Networks (ANN) and Multi-gene Genetic Programming (MGGP) models were developed. MGGP demonstrates high predictive accuracy and interpretability, while ANN excels in predicting long-term strength. Durability tests confirm the materials' applicability in brick manufacture and flexible pavement sub-base. The paper highlights the environmental and economic benefits of FA and GGBS usage, including CO₂ emissions reduction and waste diversion from landfills. The experimental-computational model provides a scalable path for incorporating waste-based materials into construction, supporting sustainable development. This study uniquely integrates multi-performance experimental evaluation with machine learning modelling for fly ash–GGBS–lime systems, addressing strength, durability, dimensional stability, and predictive reliability while benchmarking against Indian standards.</p>