Vocabulary is a vital factor in language development and the acquisition of reading and writing skills. Vocabulary development is the foundation of grammar, syntax and other linguistic development. Lexical richness can be measured using lexical diversity measures and lexical sophistication measure. Lexical diversity measures the variety of language production, that is, how many different words a person uses, while lexical sophistication is based on observing the usage of the rare words. The aim of this study is to determine whether lexical diversity measures and lexical sophistication measure can distinguish children of different chronological ages. The study involved 30 children aged 6;0 to 6;11 and 30 children aged 8;0 to 8;11. Multilingual Assessment Instrument for Narratives (MAIN; Gagarina et al., 2012; Croatian version: Hržica and Kuvač Kraljević, 2012) was used for obtaining narrative samples. The narrative material was transcribed using the program CLAN (MacWhinney, 2000) and four measures of vocabulary diversity: number of different words (NDW), type-token ratio (TTR), measure D, moving average type-token ratio (MATTR) and lexical sophistication measure were calculated. All data were statistically analysed in the SPSS program. The results have shown that the younger and older group of respondents differed statistically significantly on the NDW and D, on which the younger children achieved significantly lower results, while no statistically significant difference was recorded on the TTR and MATTR measures. No statistically significant difference was recorded on the lexical sophistication measure, although the younger group tended to use fewer rare words. Although these results can point to lower lexical diversity and sophistication among younger children, other measures of lexical diversity have not determined any differences. Therefore, the justification of drawing conclusions on lexical diversity and sophistication is limited. It is especially difficult to draw conclusions about lexical sophistication because the methodology for calculating this measure is not adapted to the Croatian language, but to English, which differs significantly from Croatian. The primary limitation of the research is the fact that only one type of language samples was used and its limited length, which made it impossible to reliably calculate some of the measures. Given the lack of research into the validation of these measures for the Croatian language, further research is needed that will include subjects across a larger chronological age range and different types of language patterns.