Introduction: Ultrafiltration (UF) is used to separate unbound drugs; however, non-specific binding (NSB) may be a limiting factor of this technique. Pretreatment of UF devices has been suggested to reduce NSB. Therefore, the pretreatment methodologies for UF devices were evaluated in order to test their effectiveness in reducing NSB of protease inhibitors (PIs). Methodology: Two PIs (lopinavir-LPV and ritonavir-RTV) were tested. UF devices were pretreated with ultrapure water, Tween-20 or Tween-80. To evaluate the NSB, after UF devices being pretreated, ultrafiltrate solutions containing the analytes at two concentrations (low and high) were used. Samples were quantified by LC-MS/MS. Results: UF devices pretreated with Tween-5% had the lowest NSB for both analytes. NSB values varied between 7 and 11% at low concentration 16-34% at high LPV concentration, respectively. For RTV, NSB was approximately 6% for low concentration and 18% for high concentration. Failure to completely remove Tween in UF devices could results in an overestimation of NSB. Conclusion: Pretreatment of UF device with Tween and subsequent removal proved to be effective in reducing NSB of PI.