Repeated motor nerve stimulation performed during neuromuscular monitoring enhances the evoked mechanical response of the corresponding muscle resulting in an increased twitch response. This is known as twitch potentiation or the staircase phenomenon. For neuromuscular modelling research twitch stabilisation techniques are often used to reduce the visible effect of potentiation, but such techniques are not always effective. Our objective was to model pharmacokinetic-pharmacodynamic (PK-PD) and twitch potentiation and to estimate neuromuscular block (NMB) in the presence of twitch potentiation. We combined a standard PK-PD model with a model describing the degree of twitch potentiation. The combined model was used to predict mechanomyographic twitch measurements and estimate NMB and twitch potentiation during muscle relaxation monitoring. Model parameters and prediction accuracy were compared to the standard PK-PD model with and without linear baseline correction. The PK-PD-potentiation model allows NMB to be estimated in the presence of twitch potentiation. It also accurately predicts data from twitch stabilisation, which is ignored with the standard PK-PD model. Compared to the standard PK-PD model, estimated PD parameters ec50 and gamma were found to be higher using the PK-PD-potentiation model. Compared to linear baseline correction, estimated PD parameters ke0 and ec50 were found to be higher. A PK-PD-potentiation model can estimate the degree of twitch potentiation and the degree of NMB during neuromuscular monitoring. This model leads to different PD parameter estimations than the standard PK-PD model however the differences are small enough to be unlikely to cause great concern among researchers.