Dystocias are common in dairy cows and are known to have labor, production, reproduction, welfare and economic implications. A technology that would accurately predict the onset of calving would help minimize the effects of dystocia by allowing producer intervention in a timely matter. The aim of this study was to assess the performances to predict calving using a decrease in vaginal temperature, rumination time, and lying time or an increase in number of lying bouts measured by 3 technologies. The combination of technologies allowing simultaneous measure of the variables was also tested. Forty 2 multiparous Holstein cows housed in tie-stall were fitted with a temperature logger, a rumination sensor and an accelerometer 7 ± 2 d before their expected calving date. Data collected during the last 120 h before calving were summarized per day and in 6 h periods. Differences between days and between periods were analyzed using proc GLIMMIX of SAS. Test performances (sensitivity, specificity, predictive values) were conducted to predict calving within the next 24, 12 or 6 h. Vaginal temperature were 0.3°C lower (P < 0.05) whereas rumination and lying time were respectively 41 and 52 min lower (P < 0.05) on calving day compared with the 4 d precalving. Cows had 2 more lying bouts on calving day (P < 0.05). While comparing tested variables, a decrease of vaginal temperature achieved the best performance to predict calving within the next 24, 12, and 6 h. Between those periods, the best performance was achieved for a prediction within the next 24 h with a sensitivity, specificity, positive and negative predictive values respectively of 74, 74, 51, and 89%. Combining the technologies enhanced the performance to predict calving with best results obtained by the combination of the 3 technologies for a prediction within the next 24 h (sensitivity: 77%, specificity: 77%, positive and negative predictive values: 56 and 90%). These results suggest that technologies are better at identifying events during which the cow did not calve than calving events. Therefore, a device that would be able to measure the 4 variables may not be able to accurately predict calving time but would provide insightful information for calving monitoring.