Computational thinking (CT) is a fundamental skill that needs to be developed by prospective mathematics teachers to improve problem-solving and logical reasoning. Integrating programming into mathematics learning is an effective approach to training this skill. This study aimed to design a hypothetical learning trajectory (HLT) for developing CT using Python programming on Google Colab. This study used a didactical design research (DDR) framework consisting of three stages: prospective analysis, metapedadidactic analysis, and retrospective analysis. The research participants were prospective mathematics teacher students enrolled in a computer programming course. Data were collected through observation, code artefacts, and reflective interviews. The results showed that HLT, designed in stages, improved the four main components of CT: decomposition, abstraction, pattern recognition, and algorithmic thinking. The students experienced improvements in breaking down problems, devising more efficient solutions, recognising patterns in code structures, and systematically designing algorithms. In addition, Google Colab supports learning by providing a collaborative and accessible programming environment. However, minor syntax errors and lack of attention to indentation were found. This study recommends using structured debugging strategies and project-based learning in optimizing CT development. The findings indicate that the integration of programming into the education of prospective mathematics teachers can equip them with essential CT skills to support technology-based mathematics teaching.