AbstractThe poor survival of ovarian cancer patients is linked to their high likelihood of relapse. In spite of full apparent macroscopic clearance, tumor recurrences arise from cells that are resistant to primary chemotherapy in the form of minimal residual disease (MRD). MRD exhibits distinct molecular drivers from bulk cancer and therefore necessitates alternative therapeutic strategies. However, there is a lack of 3D models that faithfully recapitulate MRD ex vivo for therapy development. This study constructs microfluidics‐based 3D microtumors to generate a clinically‐relevant model for ovarian cancer MRD. The microtumors recapitulate the non‐genetic heterogeneity of ovarian cancer, capturing the “Oxford Classic” five molecular signatures. Gene expression in the 3D microtumors aligns closely with MRD from ovarian cancer patients and features the upregulation of fatty acid metabolism genes. Finally, the MRD 3D microtumors respond to the approved fatty acid oxidation inhibitor, perhexiline, demonstrating their utility in drug discovery. This system might be used as a drug‐testing platform for the discovery of novel MRD‐specific therapies in ovarian cancer.