Pigs exposed to challenges with no prior experience change their daily feeding intake pattern. A method identifying deviations from normal feeding patterns could be used to develop a model framework to estimate individual nutrient requirements of challenged pigs fed with precision feeding systems. The objective of this study was to develop a tool for early identification of feed intake deviations in precision fed growing-finishing pigs. Feed intake measurements collected during 84 d in 126 growing–finishing pigs were used in this study. Electronic feeder systems automatically recorded the amount of feed consumed per meal. The recorded database was used to calculate the feed intake (DFI) per day of each pig. Individual feed intake dynamics were described by a univariate dynamic linear model (DLM) with Kalman filtering. The DLM is composed of a linear growth component, which allows the underlying level of the series to growth with a local growth factor. An unknown, but constant observation variance was dynamically estimated in the DLM. An optimized discount factor was used to specify the system variance. Finally, a standardized tabular two-side Cumsum (TC) was applied to the forecast errors generated by DLM to give warnings when pigs showed deviations of its normal feeding pattern. As the objective was identifying reduction on feed intake only alarm generated from the lower side of TC charts were considered. The DLM model was effective in following a feed intake trajectory for each individual pig over the growing period. In total, 22 pigs (17%) showed at least one deviation from normal feeding pattern. During the deviation period, when comparing forecast with the smoothed estimates, an average reduction of 30% on the DFI was observed. The system for monitoring the feeding behaviour of individual pigs based on a combination of a DLM and TC chart has proven to be a useful tool in modelling feed intake in pigs including forecasts of altered patterns. Thus, the proposed empirical approach has high potential to be integrated in a model used to estimate real-time nutrient requirements for pigs with deviation from normal feeding pattern.