Background　and　Objectives:　Coronary　computed　tomography　angiography　(CCTA)　derived　fractional　flow　reserve　(CT-FFR)　requires　a　maximal　hyperemic　state　to　be　modeled　by　assuming　the　total　coronary　resistance　decreased　to　a　constant　0.24　of　that　under　the　resting　state.　However,　this　assumption　neglects　the　vasodilator　capacity　of　individual　patients.　Herein,　we　proposed　a　high-fidelity　geometric　multiscale　model　(HFMM)　to　characterize　coronary　pressure　and　flow　under　the　resting　state,　seeking　to　better　predict　myocardial　ischemia　by　using　CCTA-derived　instantaneous　wave-free　ratio　(CT-iFR).　Methods:　Fifty-seven　patients　(62　lesions)　who　had　undergone　CCTA　and　were　then　referred　to　invasive　FFR　were　prospectively　enrolled.　The　coronary　microcirculation　resistance　hemodynamic　model　(RHM)　under　the　resting　condition　was　established　on　a　patient-specific　basis.　Coupled　with　a　closed-loop　geometric　multiscale　model　(CGM)　of　their　individual　coronary　circulations,　the　HFMM　model　was　established　to　non-invasively　derive　the　CT-iFR　from　CCTA　images.　Results:　With　the　invasive　FFR　being　the　reference　standard,　accuracy　of　the　obtained　CT-iFR　in　identifying　myocardial　ischemia　was　greater　than　those　of　the　CCTA　and　non-invasively　derived　CT-FFR　(90.32%　vs.　79.03%　vs.　84.3%).　The　overall　computational　time　of　CT-iFR　was　61　±　6　min,　faster　than　that　of　the　CT-FFR　(8　h).　The　sensitivity,　specificity,　positive　predictive　value,　and　negative　predictive　value　of　the　CT-iFR　in　discriminating　an　invasive　FFR　＞　0.8　were　78%　(95%　CI:　40–97%),　92%　(95%　CI:　82–98%),　64%　(95%　CI:　39–83%),　and　96%　(95%　CI:88–99%),　respectively.　Conclusions:　A　high-fidelity　geometric　multiscale　hemodynamic　model　was　developed　for　rapid　and　accurate　estimation　of　CT-iFR.　Compared　with　CT-FFR,　CT-iFR　is　of　less　computational　cost　and　enables　assessment　of　tandem　lesions.　©　2023　Elsevier　B.V.