The present work aims at exploring spatial thinking of primary students while they construct a robot with Lego WeDo 2.0 in groups of two. Towards this aim, content analysis of the videorecording of this procedure is performed on the basis of two proposed coding frameworks, i.e., for the spatial speech and the students’ hands gestures/movements, respectively. The results revealed varying enacted and spoken manifestations of the students’ spatial thinking and language upon which the potentiality of their scaffolding towards their further enhancement within the broader context of STEM education is revealed. Further implementations of the proposed approach at different contexts may justify the potentiality of exploration of spatial thinking during educational robotics construction tasks, providing additional domain of analysis for the enactment of spatial thinking and language expression.