This study aims at investigating the relationship between working memory updating and working memory capacity in preschool children. A sample of 176 preschoolers (36–74 months) was administered a working memory updating task (Magic House) along with three working memory capacity tests that specifically measure their core attentional component (M capacity, as defined in the theory of constructive operators): Backward Word Span, Mr. Cucumber, and Direction Following Task. Correlational analyses and cross-classification prediction analyses were performed. Updating and capacity were significantly correlated, although the correlations were not high when age was partialled out. Capacity increased with age, and mediated the relation between age and updating. More importantly, cross-classification prediction analysis revealed that high updating scores with low M capacity, and low updating scores with relatively high M capacity, are possible events; the only combination ruled out was a low updating score with precocious development of M capacity. These facts demonstrate that updating skills in preschoolers depends on M capacity but does not coincide with it. Therefore, in cognitive developmental theories, the constructs of working memory updating and capacity should be distinguished, and on practical grounds, different tests should be used to measure them.