What's a "methodological approach" for research (Master's thesis)?

I’m not super familiar with the specific field you’re researching, but I would say that there are three general approaches to methodology in the IT space: quantitative/scientific, qualitative, and artefact-oriented. Each of these general approaches, in turn, would have specific methodologies encompassed within them (for instance the Design Science Research Methodology within artefact-oriented research).

With quantitative/scientific, you’re starting off with one or more hypotheses, gathering numerical data, and then using that data to perform statistical analyses to try to disprove your hypotheses.

With a qualitative approach, you’re taking an exploratory approach to gather categorical data, usually from humans, and then examining it to see if you can spot any patterns within it, without performing any statistical analysis. This might take the form of examining user behaviour interacting with a computer program, or analysis of the contents of social media posts.

With an artefact-oriented approach, you’re looking at a problem and existing methods of solving it, designing a new method that you think might be better (or applying a method from a different area to the problem), building an implementation of this new method, and then evaluating its performance.

A methodological approach may refer to way in which a researcher intends to carry out his/her research from the specified or known methods in a discipline

You can't really choose your topic definitively until you have "mastered" the related material, but you can give a strong, well-informed description of what your topic might be.

A methodological approach is the approach you will take to exploring your topic. For example: "I will perform neural network simulations of the PFC and BG of a decision making agent while recording neuronal activity and behavior. I will supplement this computational approach with validation of behavioral results in humans." The approach should be carefully thought out (as the above example was not) and worded to make it sound nontrivial yet feasible within the scope of your program.