Estimation competence is assumed to be a relevant skill that needs to be addressed in the mathematics classroom (Sowder, 1992; Bright, 1976). However, little is known about cross-country differences in primary school students’ length estimation competence and what impact different opportunities to learn in the different educational contexts have on this competence. Taking into account an extensive analysis of estimation situations, Heinze et al. (2018) proposed a model that holds 72 estimation situations as a basis for assessing length estimation competence. Based on this model, a test on length estimation competence was developed and used in 43 German and Taiwanese classes. In total, 903 third and fourth grade students fulfilled the length estimation tasks. The test on students’ length estimation competence proved to be partially measurement invariant. Cross-country comparison revealed that Taiwanese students outperform the German students in length estimation.