Background <p>To identify which factors influence the traction time of impacted maxillary canines using a combined orthodontic-surgical approach and to create a multivariate model to predict canine traction time.</p> Methods <p>Patients undergoing traction of impacted maxillary canines by a single orthodontist were selected. Seventy patients (mean age ± SD): 19.32 ± 8.81) meeting the inclusion criteria were enrolled in the study. They had 89 impacted canines (69 palatally displaced and 20 in a buccal position). Variables that could be related to orthodontic treatment as a whole and to impaction and its treatment were recorded for each patient. The treatment predictors were explored by univariate Cox regression and multivariate generalized linear mixed models (GLMM). A multivariate analysis was performed using a mixed-effects parametric survival analysis (MMPSA) Weibull model to assess the contribution of each of the predictive variables and to create a multiple linear regression model to explain canine traction time.</p> Results <p>Cox regression showed significant association between traction time and age (<i>p</i> = 0.013) and the number of failed appointments (<i>p</i> &lt; 0.001). An MMPSA Weibull model to build a predictive model based on the Akaike information criterion (AIC) showed that interceptive maxillary expansion significantly reduced traction time (<i>p</i> = 0.002). In contrast, traction time increased with a wider angle of the impacted canine relative to the midline, with missed appointments, and with radiological overlap with lateral or central incisors.</p> Conclusions <p>The orthodontic treatment time to align impacted maxillary canines increases with radiological overlap with lateral and central incisors, a wider angle to the midline, the absence of interceptive maxillary expansion, and missed appointments.</p>