This article proposes a two-stage approach, based on the application of an analytic network process-mixed integer multi-objective programming (ANP-MIMOP) model, to solve the problem of partner selection in agile supply chains (ASCs). A key requirement of an ASC is that its constituents (suppliers, producers, distributors, etc.) can combine and react to fast-changing customer demand as efficiently and effectively as possible. An ASC needs to adopt the most appropriate supply chain structure and assign the most suitable order quantities to the most appropriate supply partners in any given circumstance. In the first stage, an ANP methodology is applied to calculate the priorities of different criteria for partner selection. Secondly, using these priorities, a MIMOP method is used to determine the supply chain structure and optimise the allocation of order quantities. An illustrative example is used to demonstrate the approach, obtain insights into its application and identify issues for future research.