This study uses the Plan-Do-Check-Act (PDCA) cycle as the hypothetical model. Multigroup analysis with public and private kindergartens as the group variables is conducted via this hypothetical model to explore the differences in its goodness-of-fit between public and private kindergartens. The research samples are selected from kindergartens in counties and cities in Taiwan, Penghu, Kinmen, and Matsu and include principals, directors (administrative personnel), teachers, and preschool educators, selected with proportional stratified random sampling. The sample structure of the formal questionnaire is as follows: 253 public kindergartens, accounting for 32.56%; and 524 private kindergartens, accounting for 67.44%. This study therefore finds that (1) the goodness-of-fit of the overall model for both public and private kindergartens is good; (2) the results of a measurement invariance test demonstrate measurement constancy and cross-group validity; (3) regarding measurement invariance in the public and private kindergartens, the structural model shows measurement constancy and cross-group validity; (4) for “private” kindergartens, the impact of “Do” on “Action” is higher than for “public” kindergartens, and there is no constancy or cross-group validity. Accordingly, this study advocates the following: (1) subsidizing the software and hardware construction and personnel costs of private kindergartens to improve the quality of kindergarten education; (2) promoting relevant policies such as the “public-private cooperation model” to emphasize the decommodification of early childhood education (ECE); and (3) implementing the "community-orientation" and "localization" of kindergartens to improve and strengthen their connections with families and communities.