Background: Diffuse myocardial fibrosis and microvascular dysfunction are suggested to underlie cardiac dysfunction in patients with type 2 diabetes, but studies investigating their relative impact are lacking. We aimed to study imaging biomarkers of these and hypothesized that fibrosis and microvascular dysfunction would affect different phases of left ventricular (LV) diastole. Methods: In this cross-sectional study myocardial blood flow (MBF) at rest and adenosine-stress and perfusion reserve (MPR), as well as extracellular volume fraction (ECV), were determined with cardiovascular magnetic resonance (CMR) imaging in 205 patients with type 2 diabetes and 25 controls. Diastolic parameters included echocardiography-determined lateral e’ and average E/e’, and CMR-determined (rest and chronotropic-stress) LV early peak filling rate (ePFR), LV peak diastolic strain rate (PDSR), and left atrial (LA) volume changes. Results: In multivariable analysis adjusted for possible confounders including each other (ECV for blood flow and vice versa), a 10% increase of ECV was independently associated with ePFR/EDV (rest: β = − 4.0%, stress: β = − 7.9%), LA_max /BSA (rest: β = 4.8%, stress: β = 5.8%), and circumferential (β = − 4.1%) and radial PDSR (β = 0.07%/sec). A 10% stress MBF increase was associated with lateral e′ (β = 1.4%) and average E/e’ (β = − 1.4%) and a 10% MPR increase to lateral e′ (β = 2.7%), and average E/e’ (β = − 2.8%). For all the above, p < 0.05. No associations were found with longitudinal PDSR or left atrial total emptying fraction. Conclusion: In patients with type 2 diabetes, imaging biomarkers of microvascular dysfunction and diffuse fibrosis impacts diastolic dysfunction independently of each other. Microvascular dysfunction primarily affects early left ventricular relaxation. Diffuse fibrosis primarily affects diastasis. Trial registrationhttps://www.clinicaltrials.gov. Unique identifier: NCT02684331. Date of registration: February 18, 2016.