This study investigates the relationship between mental-state language and theory of mind in primary school children. The participants were 110 primary school students (mean age = 9 years and 7 months; SD = 12.7 months). They were evenly divided by gender and belonged to two age groups (8- and 10-year-olds). Linguistic, metacognitive and cognitive measures were used to assess the following competencies: verbal ability, use of mental-state terms, understanding of metacognitive language, understanding of second-order false beliefs, and emotion comprehension. Correlations between children' use of mental-state language and their performance on theory-of-mind tasks were moderate, whereas correlations between children's comprehension of such language and ToM abilities were high. In addition, regression analyses showed that comprehension of metacognitive language was the variable which best explained children's performance on both false belief tasks and an emotion comprehension test when verbal ability and age were controlled for. (Contains 4 tables.)