Objectives. To assess the impact of daily life‐stress on behaviours (alcohol consumption, food choice and physical exercise) that influence health risk.Method. Twenty‐three nurses and 21 teachers completed daily assessments of mood, alcohol and food intake, together with weekly measures of hassles, perceived stress, anxiety, depression and exercise behaviour, over an eight‐week period. Comparisons were made between the two highest and two lowest perceived stress weeks.Results. Work and home‐related hassles were more frequent on high than low stress weeks, and anxiety and depression levels were greater, even after controlling for negative affectivity. Participants who drank alcohol to cope showed increases in consumption between low and high stress weeks. ‘Fast food’ was eaten more frequently during high stress weeks. People who reported that mood control was an important influence on their choice of food showed increases in the amount of cheese they ate, and greater consumption of sweet foods. There were no differences in exercise frequency or duration with changes in perceived stress.Conclusions. Diary studies allow changes in health behaviour with life‐stress to be assessed naturalistically. The impact of stress depends on the moderating effects of motives and coping expectancies associated with specific behavioural domains. Some responses may be prejudicial to long‐t