The principles of pharmacokinetic (PK) dose tailoring in clinical practice, using limited blood sampling and Bayesian PK analysis, have been described for factor VIII (FVIII). This study applied the same procedure to recombinant FIX (rFIX), i.e. population PK modelling and the use of a simplified (one-compartment) model to describe only the terminal part of the coagulation factor vs. time curve. Data from a previous study on rFIX in 56 patients (4–56 years, 18–133 kg) were used to define a three-compartment population PK model. The average FIX clearance was 8.4 mL h−1 kg−1. Elimination half-life ranged between 14 and 27 h. Data obtained from 24 h after the infusion were found to define the terminal phase of FIX disposition. Doses to produce a target trough FIX level (set at 0.01 IU mL−1) at 72 h predicted by the Bayesian analysis, with blood sampling at either 24, 48 and 72 h or at only 24 and 48 h, were within −40% to +67% of those predicted using the three-compartment model, and within −57% to +125% for targeting a level at 96 h. These errors were lower than the overall interindividual variance in dose requirements. As three-compartment models are needed to characterize the PK of both plasma-derived FIX and rFIX, simplification to a one-compartment model is less straightforward than for FVIII, and the methodology should be investigated further before clinical application. Limited blood sampling and Bayesian analysis could still, however, be potentially useful for targeting rFIX trough levels during prophylaxis.