Classical subjects and competencies, such as algebraic thinking, have a recognizable categorization. Although these categorizations may have small variations between educational systems, levels of education or fields of application, in the curricula of any country or university there are great similarities between them. This is not the case with computational thinking. Since Wing gave it a boost in 2006, this new competence has been gaining ground in the teaching practice of different subjects and it is currently being incorporated into the curricula of a multitude of educational systems. Learning computational thinking involves working on concepts such as logical reasoning, algorithmic thinking and problem-solving techniques. However, its definition is not yet unique, although affinities are being achieved in this sense. But its categorization is a point on which positions have not been brought closer. Since this competency is not usually an independent subject, but is linked to one or more subjects, its categorization usually depends on this or these subjects. In this paper, we study different categorizations of computational thinking and make a proposal agreed upon by several entities from different countries. Categorization of computational thinking refers to the identification and classification of its key components and skills. This categorization helps to better understand how computational thinking can be taught, learned and applied in different contexts. The categorization of computational thinking is crucial to its teaching and learning. By identifying and categorizing these skills, educators can develop more effective curricula and students can better understand how to apply these skills in different contexts. In this paper, by qualitatively analyzing the different categorizations, and studying their implementations in different types of subjects through a content analysis, we obtain as a result a global categorization.