Rheumatoid arthritis (RA) is a major public health concern, which can cause serious outcomes. Low-dose methotrexate (MTX) is a cornerstone in RA treatment, but there is significant heterogeneity in clinical response. To evaluate underlying sources of pharmacokinetic variability and clinical response of MTX, a physiologically based pharmacokinetic and pharmacodynamic (PBPK/PD) model was developed using PK-sim and Mobi (version 11.1). The PBPK model included metabolism and transportation by AXO1, FPGS, GGH, RFC, and MRP2, with renal and biliary excretion. We also developed various degrees of renal insufficiency populations with subsequent dosing optimizations. A total of 23 MTX plasma concentration-time profiles were used, with 97% of predicted plasma concentrations within a two-fold range compared to observed data. The PBPK/PD modeling and simulation demonstrated that variability in renal clearance and enzymes related to MTX are likely important drivers of PK variability and there is a quantitative relationship between MTX-PG3 and RA treatment response. The PBPK/PD model could be used to guide improvement in MTX dose regimens for RA patients.