Objective: to construct and validate an instrument to assess talent retention in nurses. Methods: methodological study whose stages were as follows: construction of the instrument by integrative literature review; content validation; and agreement by experts. To establish psychometric properties, it was applied to 290 nurses for construct validity through exploratory factor analysis, principal components method and Promax rotation. Evaluation of the internal consistency of the constructs by Cronbach's Alpha coefficient and McDonald's omega coefficient. Results: content validation by 10 experts obtained a degree of agreement >80% and second-order agreement coefficient (AC2 Gwet) between 0.83 and 0.89. After factor analysis, we obtained: instrument with 25 items, high degree of reliability and internal consistency; Kaiser-Meyer-Olkin test = 0.947; Bartlett p<0.001; and α from 0.87 to 0.92 with the formation of four factors explaining 74% of the variance. Conclusion: the instrument showed evidence of content and construct validity and adequate internal consistency. Contributions to practice: this innovative instrument for Nursing science enables managers to identify their teams' own talent retention factors, directing assertive measures for retaining nurses according to the predominant motivators.