The aim of this qualitative research is to describe the thinking process as a profile of students' mathematical computational thinking (CT) skills in terms of CT attitudes. The subjects in this study were 66 junior high school students in grade IX. There were three students taken by purposive sampling based on high, medium, and low CT attitude. The main research instrument was the researcher, while the auxiliary instruments were mathematical CT ability test and CT attitude scale questionnaire. Data analysis included 3 stages, namely data reduction, data presentation, and conclusion drawing. The data credibility test used triangulation techniques, namely tests and interviews. The results showed that subjects with high CT attitude were able to show all indicators of CT ability, namely abstraction, problem decomposition, algorithmic thinking and generalisation and pattern recognition. However, subjects with moderate and low CT attitudes were only able to show one of these indicators, namely abstraction. Based on the results of this study, researchers will create a mathematics learning model to develop CT skills, especially for number pattern material.