Inadvertent perioperative hypothermia, defined as a body temperature <36.0°C, is a common outcome of anesthesia that can cause serious consequences to patients. The aim of this study is to determine the prevalence of inadvertent hypothermia among surgical procedures from two referral centers in Brazil and to identify sociodemographic, clinical, or surgery-related predictors of hypothermia. This is a cross-sectional study, conducted at two public hospitals in Brasília, Brazil. After the exclusion of 109 patients, 312 subjects (American Society of Anesthesiologists [ASA] physical status I-III) were enrolled from July 2016 through July 2018. The main outcome measures were the prevalence of hypothermia and its predictors. The mean age of the 312 patients was 43.2 (18.2) years (range 18-85 years), and 186 (59.6%) were female. The prevalence of inadvertent hypothermia was 56.7%. Predictors of hypothermia were perioperative chills ( p = 0.026), patient's body temperature on arrival in the operating room ( p < 0.001), diabetes ( p < 0.001), ASA status III ( p < 0.001), systolic blood pressure ( p < 0.001), general anesthesia ( p < 0.001), medical specialty ( p < 0.001), fentanyl-based anesthesia ( p = 0.002), and surgery time ( p < 0.001). The multivariable model prediction model for hypothermia showed fairly good discrimination (area under the receiver operating characteristic: 79.0%, 95% confidence interval 68.0 to 80.1). Approximately 6 in 10 patients undergoing surgery developed inadvertent perioperative hypothermia. The risk of hypothermia is influenced by a myriad of factors that can be used in simple and low-cost predictive models with adequate discriminatory power.