To enable the high milk production energy-dense food needs to be fed to high-producing dairy cows. This kind of food consists of easily fermentable carbohydrates at the expense of fibre. Thus,this feeding management doesn’t meet the physiological requirements of ruminants and therefore enhances the risk of health problems like subclinical rumen acidosis (SARA) and its conse-quences. During SARA-conditions the eating and ruminating behavior, which is associated with rumen health, is changing. Automatic ruminating measuring systems allow conclusions of rumen health and enable replacement of time-consuming observations of dairy cows. The objective of this study was to perform a validation of the system “Rumiwatch” (Itin + Hoch GmbH, Switzerland) during different feeding strategies in cows. For this purpose 216 observations of 10 minutes were conducted on eight dairy cows. These observations were divided into three phases (baseline with-out concentrate, adaptation with increasing concentrate concentration und SARA-challenge with 65 % concentrate). Subsequently, the visual observations were compared to the automatically stored data. The statistical analysis illustrate the decrease of ruminate time, ruminate chews, chews per mi-nute, chews per bolus and the number of boluses with increasing portion of concentrate (P < 0.01). The rumination time did not differ among both measuring systems (P = 0.58), whereas the number of ruminate chews, the number of chews per minute as well as per bolus were lower in visual observations comparing to automatically observation (P?0.02). Even the number of boluses showed significantly differing values by using the automatic measuring method compared to the visual observation (P < 0.01). However, no interaction between the feeding phase and the meas-uring method was found, suggesting that feeding effects are negligible regarding the agreementof the measuring methods. Furthermore, the correlation analysis resulted in good relations be-tween visual and automatic recorded values. Thus a very high correlation coefficient was deter-mined for ruminate time (r = 0.92), ruminate chews (r = 0.96) and number of boluses (r = 0.91), a high coefficient for chews per bolus (r = 0.86) and a moderate correlation for chews per minute (r = 0.63) caused by an outlier. It can be concluded that the Rumiwatch-system can replace time-consuming observations by dif-ferent feeding strategies und can be used to get hints of modified chewing activity in dairy cows.