Based　on　the　calculation　methods　of　equivalent　water-cement　ratio　of　concrete　and　average　paste　thickness　of　aggregates　provided　by　authors,　the　nonlinear　relation　between　28　d　concrete　compressive　strength　and　equivalent　water-cement　ratio　of　concrete,　average　paste　thickness　of　aggregates,　fly　ash-binder　ratio　is　established　by　using　artificial　neural　network.　The　outcome　can　be　used　to　reduce　the　number　of　trial　and　error　and　to　save　cost,　labor　and　time　in　concrete　mix　proportion　design,　and　can　further　lay　the　foundation　for　the　mix　proportion　design　of　high　volume-stability　concrete.