Few studies have focused in calf behavior as an early predictor of health disorders. The objective of this study is to use calf behavior as early predictor of health disorders. A total of 325 calves (male calves from 30 to 90 d of life) were fitted with pedometers (Trackacow, ENGS system, Israel) to measure calves steps, lying time, standing time, changing position from lying to standing (swaps) and access to the feed bunk (number of times and time per access). The incidence of health problems, weaning time and preventive treatment time (vaccine and antibiotics in water) were recorded. Data were analyzed by using GLIMMIX procedure of SAS (version 9.4) and significance was declared at P < 0.05. On average, calves spent 16.3 h/d lying, 4.7 h/d standing, changed position from lying to standing 19.3 times/d, and did 1,468 steps/d. They attended to the feed bunk 10.4 times/d and spent a total of 3.0 h/d (average of 17.3 min/visit). Sick calves (n = 24, 80% respiratory diseases) versus healthy calves spent more time lying (+5.0%), less time eating (−5%) and did less steps (−5.0%) during 3 to 4 d before being diagnosed. These changes in behavior allow predicting the occurrence of health disorder with 73% accuracy 3d before actual diagnostic. Moreover, weaning increased eating time (+15%), the number of visits to the feed bunk (+13%) and the number of steps (+7.0%); and preventive treatments reduced intake time (−20%), the number of visits to the feed bunk (−5%) and the number of steps (−11%), and increased the number of swaps (+9%). These effects have to be considered in the interpretation of results to avoid false alarms. Calf behavior monitored with pedometer can help to predict sick calves before they manifest clinical symptoms allowing early treatment.