This study aimed to assess the agreement between the total energy expenditure (TEE) estimated by the activPAL® triaxial accelerometers (ACC) and the TEE measured by the doubly labeled water method (DLW), as well as to assess if these values differ between the classifications of body mass index (BMI). This is a cross-sectional study. Low-income adult women (19-45y) with BMI ≥ 18.5 kg/m2 were included. Accelerometry data (activPAL® ) were collected over 7 consecutive days, which were used to calculate TEE-ACC and compared with DLW data. The Bland-Altman method, concordance correlation coefficient and root mean square error were used to assess agreement between methods. The sample consisted of 55 women with a mean age of 31 ± 5 years. The agreement between TEE-ACC and TEE-DLW showed a bias of -142.5 kcal (-7.1%). Among the BMI classifications, participants with normal weight show a bias of -417.1 kcal (-21.0%), participants with overweight, -87.5 kcal (-3.9%) and participants with obesity, 97.5 kcal (4.3%). Furthermore, the bias between the methods showed a significant and positive correlation with the body weight (r = 0.49; p < 0.01). The TEE-ACC estimates from activPAL® were reasonably accurate when compared to the TEE-DLW, especially in women with overweight and obesity, being much less accurate in individuals with normal weight.