This study delivers a Michigan-specific mechanistic-empirical framework for rehabilitation pavement design by locally calibrating PMED V2.6 across all major MDOT rehabilitation fixes, including asphalt over asphalt/concrete/rubblized concrete/crushed asphalt, and concrete over asphalt/concrete. Using 14 field projects, FWD, GPR, LWD, and field cores data were evaluated. The PMED materials inputs were established for existing aggregate base, sand subbase, pavements, crushed asphalt, rubblized concrete, and ASCRL asphalt mix. Calibration employed 336 pavement sections (308 asphalt, 28 concrete) and included fatigue cracking (top/bottom), rutting, transverse cracking (thermal + reflective), faulting, and IRI. Bootstrapping and reliability analyses produced revised local coefficients and standard-error equations that reduce bias relative to national defaults. Validation against AASHTO 93 showed calibrated PMED overlay thicknesses typically within ±1 to ±1.5 inches, aligning mechanistic predictions with current practice while improving sensitivity to materials, climate, and distress interactions. The PMED limitations were identified for some rehabilitation fix types, leading to recommended alternate layer inputs or modeling approaches. Sensitivity analyses also defined PMED input constraints, including minimum HMA overlay thicknesses of 3.5 inches over existing HMA and 4.5 inches over existing concrete pavements. Key outcomes include statewide layer modulus guidance, rehab-specific performance threshold recommendations, adoption of rehabilitation Level 2 inputs for HMA over HMA, and recommended design reliability levels to effectively integrate mechanistic-empirical rehabilitation design into MDOT standard practice.