M3 = mean(Mar, 3);
M4 = mean(Apr, 3);
M5 = mean(May, 3);
^^ ^^^ These are both signs of bad data design!
The problem occurs when you decided to have lots of separate variables in the workspace, each one named with a month of the year. Remember that meta-data (e.g. month names and pseudo-indices) is data, and data should be stored in arrays and not forced into variable names. By naming your variables like that you forced yourself into writing slow, complex, inefficient, buggy code when you try to access those names dynamically (like you show in your question):
Note that copy-and-pasting code like you show is another sign that you have done something wrong: computers are really good at repeatedly doing simple tasks in loops. So when you sit and copy-and-paste code like that, you are just doing the computer's job.
In contrast if you had used matrices or arrays (e.g. a cell array, a structure array, a table) to store that data then accessing that data in a loop would be efficient and robust. The name "MATLAB" comes from "MATrix LABoratory" and not from "lets names all of the data with different names in lots of variables and make the data harder to process".
So fo you have not told us the most important information: how did you get those months into the workspace? That is the correct place to start to fix this data design: by e.g. loading them all into one array.
