In modern dairy farms, automatic monitoring systems are needed to record changes in the health status of individual cows. Modern automatic monitoring systems allow the recording of animal behaviour and include data on movement, feeding and rumination times, standing and lying time. In practice, the computer programs used to analyse data on movement activity and feed intake and rumination times from the sensors used are mostly focused only on heat detection. A suitable way of evaluating these data for early detection of the start of birth or the beginning of disease in animals is not yet usual. The results presented in the methodology show, among other things, that changes in the time of feeding and rumination can be practically used, in addition to refining the peak of estrus, also to detect the approaching birth and the occurrence of certain types of diseases. The presented methodology therefore extends the possibilities of applying records of cow feeding behaviour to new procedures usable in optimising dairy herd management. The methodology is addressed to dairy cattle breeders, expert advisors and consultants and developers of computer programs for herd management that evaluate data from sensors recording the movement activity and feeding and rumination time of heifers and cows.