Goldin-Meadow et al. (2008) examined whether word orders of speakers’ own languages influence their non-verbal behaviors by asking speakers of English, Chinese, Spanish, and Turkish to describe events non-verbally using gestures. They found that speakers of all four languages preferred to perform gestures in Actor-Patient-Action order. Although they argued that this reflects a natural order of event description for humans, their research is limited in languages with Subject-Object (SOV or SVO) word order. In order to verify their claim, we examined Kaqchikel, a Mayan language spoken in Guatemala, employing Object-Subject (VOS) word order. In our non-verbal event description experiment, 32 Kaqchikel native speakers described 18 pictures of transitive events using gestures. We found predominant Actor-Patient order (86.2%) as well as Patient-Actor order (13.8%). These results revealed that event descriptions are not only governed by universal conceptual preference to Actor-Patient order but also influenced by word orders of speakers’ own languages.