The number of automatic milking systems (AMSs) installed worldwide shows an increasing trend. In comparison to the preliminary models, new versions employ more sophisticated sensor technology than ever before. The originally developed AMSs were characterised by larger vacuum fluctuations and vacuum reductions than conventional milking systems. The objective of this study was to find out whether this situation still holds or if an improvement has occurred. The vacuum behaviour at the teat end of an artificial teat during simulated milking was measured in a study that involved different AMS types (AMS A, B and C). Each system was tested over a range of flow rates (0.8 to 8.0 L/min). The wet-test method was used and teat-end vacuum behaviour was recorded. At a flow rate of 4.8 L/min, the lowest vacuum fluctuation (6.4 kPa in b-phase) was recorded for AMS A, while the lowest vacuum reduction (3.5 kPa in the b-phase) was obtained for AMS B. AMS C yielded higher values for vacuum reduction and vacuum fluctuation. Consequently, it was concluded that AMS A and B, in terms of construction and operational setting (vacuum level), are more appropriate than AMS C. Nevertheless, high values for vacuum reduction or fluctuation have a negative effect on the teat tissue. Hence, one of the future challenges in milk science is to develop a control system that is able to allow fine adjustments to the vacuum curve at the teat end.