Dynamic whole-body (DWB) PET data for irreversible tracers, such as [ 18 F]-FDG, can be obtained with conventional PET scanners using multi-bed multi-pass data acquisition protocols, providing parametric images, which are more informative than standard SUV images. The drawback is a relatively long scanning time. Recently, a dual injection protocol with a reduced scanning time was proposed for a total-body PET scanner (Wu et al., JNM 2022). We developed a dual-injection protocol for a conventional PET scanner with multi-bed multi-pass acquisition. The input function is derived from the heart or aorta, and fitted with an analytical function, with the 1 st part being obtained after the 2 nd injection. Two different models were used for fitting the time-activity curves: standard compartmental modelling (CM) and a combination of CM and Patlak analysis. We have evaluated the protocol using computer simulations, and investigated the effect of different injection fractions, injection time-points, scan start-times as well as different values of blood volume (V b ) and tracer non-irreversibility (k 4 ). Our results showed that the injection fraction and time point for the 2 nd injection had minor impact on the estimated parameters. While a late scan start time as well as non-zero V b or k 4 values could result in bias. Both bias and variance were lower with the combined model as compared to the standard model. Also, the parameter estimation was not very sensitive to errors in injection fraction with the combined model. In conclusion, we found that the proposed protocol is feasible for obtaining dynamic whole-body PET data with irreversible tracers, using the multi-bed multi-pass acquisition protocol, while still measuring individual AIFs.