Background: The relation between mechanical ventilation (MV) and bronchopulmonary dysplasia (BPD)-a common disease in extremely premature newborn (PTNB)-is well stabilished, but is unknown, however, how much time under MV influences the severity of the disease. Aim: To define the duration under MV with greater chance to develop moderate to severe BPD in extremely PTNB and to compare clinical outcomes before and during hospitalization among patients with mild and moderate to severe BPD. Methods: Fifty-three PTNB were separated into mild and moderate to severe BPD groups and their data were analyzed. Time under MV with a greater chance of developing moderate to severe BPD was estimated by the ROC curve. Perinatal and hospitalization outcomes were compared between groups. A logistic regression was performed to verify the influence of variables associated to moderate to severe BPD development, such as pulmonary hypertension (PH), gender, gestational age (GA) and weight at birth, as well the time under MV found with ROC curve. The result of ROC curve was validated using an independent sample (n = 16) by Chi-square test. Results: Time under MV related to a greater chance of developing moderate to severe BPD was 36 days. Moderate to severe BPD group had more males (14 vs 5, p = 0,047), longer time under MV (43 vs 19 days, p < 0,001), more individuals with PH (12 vs 3, p = 0,016), worse retinopathy of prematurity (grade 3, 2 vs 11, p = 0,003), longer hospital length of stay (109 vs 81,5 days, p < 0,001), greater PMA (41 vs 38 weeks, p < 0,001) and weight (2620 vs 2031 g, p < 0,001) at discharge and the mild BPD group had more CPAP use prior to MV (12 vs 7, p = 0,043). Among all variables included in logistic regression, only PH and MV < 36 days were significant in the model, explaining 72% of variation in moderate to severe BPD development. In the validation sample, prevalence of preterm infants who needed MV for more than 36 days in the moderate to severe BPD group was 100% (n = 6) and 0% in mild BPD group (p = 0,0001).