Process industry remains one of the difficult-to-decarbonise sectors globally. To mitigate industrial greenhouse gas (GHGs) emissions, an eco-industrial energy systems (e-IES) optimisation framework is proposed by coupling mathematical optimisation with clustering algorithms and first principle modelling. Within the framework, a rooftop farming database was developed using biogeochemical simulations, which models seven crop growth in response to 10 cultivation conditions. Clustering algorithm was applied to analyse energy system data, along with the rooftop farming database, to inform the optimisation model. A Mixed Integer Linear Programming optimisation model was developed to optimize system design considering the trade-off between economic and environmental objectives. The implications of rooftop design on e-IES and their interactive effects on industrial decarbonisation were addressed. A case study at an industrial park in Suzhou China reveals that rooftop farming could generate mutual benefits from both cost and GHG reduction perspectives. Planting lettuce indicates a cost-efficient solution, and planting tomato could contribute the most to GHG emission reduction. Compared to the rooftop PV and the spare rooftop, 2.4% and 5.6% cost savings, as well as 10.2% and 16.3% emission savings, could be achieved respectively by implementing rooftop farming. Overall, this study demonstrates an emerging perspective on decarbonising the industrial sector by coupling biogeochemical simulation and energy system optimisation and adopting cross-disciplinary approaches.