This study investigated the potential for accurate detection of clinical mastitis (CM) in an automatic milking system (AMS) using electronic data from the support software. Data from 358 cows were used to develop the model which was then tested on 2 independent data sets; one with 311 cows (same farm different year) and one with 568 cows (from a different farm). Data from a common period was captured for healthy cows (n = 1066), single quarter (n = 101) and multi-quarter (n = 70) CM cows. Clinical mastitis was determined by visual inspection of suspect quarters for presence of redness, heat, swelling with flakes/clots in the milk, having prior elevated electrical conductivity (mS/cm, EC), and treated with antibiotic. All data was assessed using logistic mixed models. Twelve parameters were included in the initial model before a backward elimination which resulted in the following 6 parameters being included in the final model: quarter level milk yield (kg, MY), EC, average milk flow (kg/min), and occurrence of incompletely-milked quarters in each milking session; MY per hour and EC per hour between successive milking sessions. All measurements were assessed to determine their ability to detect CM, both as individual variables, but also as combinations of the 12 abovementioned variables. These were assessed by producing a receiver operating characteristic curve and calculating the area under the curve (AUC) for each model. Overall, 6 measurements of final model had significant (P < 0.05) mastitis detection ability as separate predictors. The best mastitis prediction was possible by incorporating 6 measurements as well as the random cow/quarter effects in the model resulted in sensitivity (Se) of 90%, specificity (Sp) of 91% and AUC (0.96). Assessment of the model found robust results with sound Se (90 to 100%), Sp (87% to 96%) and excellent AUC (>0.9). This study demonstrated that improved mastitis status prediction can be achieved by using multiple measurements and any new index based on multiple measurements is expected to result in improved accuracy of mastitis alerts thereby improving the detection ability and practicality on farm.