This study explored how early childhood assessment policies are implemented in the context of US Head Start programs. To achieve this, A Head Start and B Public School Head Start programs were selected as case studies, and in-depth interviews were conducted with four teachers and two educational administrators who supervised the teachers. Classroom observations were also conducted. Qualitative data analysis was performed on the basis of the research questions, with attention paid to the distinct characteristics seen in each case. The results showed that the two programs used contrasting evaluation tools: A Head Start used criterion-referenced assessments that are based on teacher observations, while B Public School Head Start used standardized direct assessment methods. In addition, the early childhood assessments that were implemented at the classroom level showed a close linked to the organization and operations of small group activities in each classroom. Both institutions made efforts to use the assessment results for improvements at both the classroom and institutional levels; however, the ways in which the results were utilized varied, as did their emphasis. Finally, the practices of early childhood assessments differed with the institutions’ approaches to curriculum implementation and the relationships among their staff. The findings of this study are important, in that they explore how early childhood assessment policies can be applied in practice, examining the diversity of their implementation in the context of curriculum operations and the relational contexts of two types of Head Start program.