This paper investigates whether deciduous upper molars and lower canines have sexual dimorphic features, exploring these teeth' dimensions and the presence of Zuckerkandl’s tubercle and Carabelli’s cusp on the first and second upper molars. We analyzed 64 pairs of dental plaster casts from 34 females and 30 males aged between 3 and 12 years. We measured the first and second deciduous upper molars and the lower deciduous canines (maximum mesiodistal and buccolingual length), and we registered the presence of the Zuckerkandl’s tubercle and the Carabelli’s cusp on the first and second upper molars, respectively. Regarding the differentiation between sexes using Carabelli’s cusp and Zuckerkandl’s tubercle, the classification was not independent of Carabelli’s cusp presence only for tooth 65 (p = 0.035). In all other teeth, whether for Carabelli’s cusp or Zuckerkandl’s tubercle, their presence was similar for both sexes. There were statistically significant differences between sexes (p < 0.05) for the buccolingual measurements of both upper second molars, the first right upper molar, and the right canine. The developed model allowed for a 64.1% accuracy in sex estimation. The study suggests that while Carabelli’s cusp and Zuckerkandl’s tubercle in upper deciduous molars don't consistently differ between sexes, tooth size, particularly the buccolingual measurements of certain teeth, including upper deciduous molars and lower canines, may provide a more reliable criterion for sex estimation. The developed model depicted moderate accuracy, underscoring the need for a multifactorial approach when estimating sex from skeletal remains. It suggests that while dental features can contribute to sex estimation, they should be used in conjunction with other skeletal or molecular markers to improve accuracy. •Carabelli’s cusp in the upper left second molar is a sexual dimorphic feature.•Zuckerkandl’s tubercle in the first upper molar doesn’t show sexual dimorphism.•The buccolingual dimension shows more sexual dimorphism than the mesiodistal one.•All measurements in the upper second deciduous molar showed sexual dimorphism.•Our model allowed for 64.1% of correct sex classifications.