There has been substantial research in recent years on the development dynamics of agglomeration economies - the industrial cluster perspective. However, the industrial cluster literature has tended to neglect the environmental impacts of clustering. Meanwhile, industrial ecology has tackled environmental concerns by promoting a new type of eco-friendly industrial system, eco-industrial park (EIP), that encourage circular flows of materials, energy and by-product exchange between neighbouring firms. Given that these two theories can potentially supplement each other, this doctoral study seeks to build bridges between the industrial cluster literature and the industrial ecology literature in order to address the research question: "how do clustered producers draw on inter-firm collaboration and institutional linkages to undertake holistic environmental upgrading strategies?" This study draws on comparative empirical evidence from the Banwol-Sihwa textiles dyeing cluster and the Yeosu petrochemical cluster in South Korea. Both the industrial clusters show a pattern of gradual environmental upgrading that is termed as 'collective eco-efficiency', although specific types of such collective behaviours have become differentiated due to their structural differences. The findings provide three key contributions to industrial ecology: (1) Intra-sectoral inter-firm relationships and regional firm-institutional relationships are much more influential in establishing an EIP than inter-sectoral inter-firm relationships on which industrial ecology has conventionally emphasised. (2) Although industrial ecology has mainly concerned of sectoral heterogeneity within an area, the multiplicity of production stages within an industrial cluster is also an important physical condition in establishing an EIP. (3) The notions of institutional setting and planning that have used in very ambiguous manners in industrial ecology are in essence a type of learning-by-interaction dynamics between local firms and other actors based on regional innovation systems. As for the industrial cluster literature, the findings in this study also suggests three conceptual implications: (1) Most studies addressing environmental issues in the industrial cluster literature have largely ignored the inter-sectoral dimension, the dynamics of collective efficiency is not necessarily limited within a sectoral boundary at least with regard to environmental upgrading. (2) Given that value chains do not necessarily match material flows, anchor tenant's coordination power over material flows should be highlighted in addition to lead firm's coordination power over value chains. (3) Regional innovation systems function as institutional ability to balance private benefits and social benefits.