The three-pass reading method
When you're handed a regex you didn't write, don't try to swallow it whole. Three passes, in order:
Pass 1: Find the anchors and groups. Look for ^, $, (...), (?:...), (?=...). These are the structural skeleton — they tell you the *shape* of the pattern.
Pass 2: Identify each character class and quantifier. What's being repeated? How many times? What set of characters? Translate each piece into English: [a-z]+ = "one or more lowercase letters."
Pass 3: Read the whole thing as one English sentence. If you can speak it, you understand it. If you can't, paste it into regex101 and read the explanation pane.
Worked example
Pattern: ^(\w+)\s+(\d{3,4})$
Pass 1 — anchors & groups: ^...$ means whole-string match. Two capturing groups separated by something.
Pass 2 — pieces: \w+ = one or more word characters. \s+ = one or more whitespace. \d{3,4} = three or four digits.
Pass 3 — sentence: "A line that is exactly: a word, then whitespace, then a 3-or-4 digit number. Capture the word as group 1 and the number as group 2."
Things to look up, not memorize
Some pieces appear rarely enough that you should look them up rather than memorize them: \p{...} Unicode categories, (?(name)yes|no) conditionals, (?>...) atomic groups. Knowing they exist is enough.