In contrast to a growing interest in the ecological implications of consistent individual differences in behavior, it is still unclear how consistent those differences are across environmental gradients, especially under ecologically relevant contexts and timescales. We investigated how individual variation in swimming activity of Eurasian perch (Perca fluviatilis) changes in response to natural fluctuations of water temperature as a proxy for an ecological gradient for ectothermic animals. Using an automated acoustic telemetry system, we tracked individual positions of adult perch in a whole natural lake for about 1 year and compared several model frameworks for the individual responses to water temperature. The best-fit model revealed that rising water temperature led to an increase in both between- and within-individual variation in activity, thereby stabilizing behavioral repeatability across temperature. Further, including temporal autocorrelation significantly improved the model performance, suggesting that consistent individual differences in behavior could be partially explained by factors such as slowly changing states in the wild. By using ecologically relevant data, our results revealed complex patterns of behavioral variation in response to an environmental change represented by water temperature.