Number sense has been identified as an important foundational skill in the development of later mathematics competence. Although number sense has historically been difficult to define in the educational literature, operational definitions of the construct typically consist of a collection of early numeracy skills or “number sense components” such as quantity discrimination, rote counting, and one-to-one correspondence. Consequently, assessments of number sense tend to measure a wide variety of these skills. The purpose of this study was to investigate the predictive validity of three measures of number sense: the Test of Early Numeracy (TEN), Number Sense Brief Screener (NSB), and Early Numeracy Test (ENT). This study also sought to identify which measure or combination of measures best predicted later mathematics achievement, as measured by the Test of Early Mathematics Ability, Third Edition (TEMA-3). Number sense assessments were administered to participants at kindergarten entry and the TEMA-3 was administered at the end of kindergarten. Data were analyzed using simple linear regression analyses, multiple regression analyses, and a procedure for comparing dependent correlations. Evidence for the predictive validity of each number sense measure was demonstrated; however, statistically, no number sense measure emerged as the best predictor of later mathematics achievement. The combination of the NSB with either the TEN or the ENT explained variation in TEMA-3 scores better than the NSB alone, but this finding may not be of clinical importance. The concurrent and predictive validities of teacher rating of student number sense were also examined. Results indicated that the TEN, NSB, and ENT all predicted TEMA-3 scores better than teacher rating of student number sense in the fall. Teacher rating of student number sense in the spring explained 42% of variation in TEMA-3 scores. Implications for practice and directions for future research are discussed.