1) Find a good source of information --- typically, this is either very good lectures (like on youtube), a good textbook, or good lecture notes.
2) Do problems. There is a fairly large gap between those that just watch the lectures and those that have sat down and try to go through each and every step of the logic, and that's what everyone here (on HN) is pointing out when they similarly mention doing problems.
2b) Have solutions to those problems. I make this a separate point because it's important to spend quality time on a problem yourself before looking at the solutions. At the end of the day, if you read the problems and then the solution right away, that's much closer to reading the textbook itself instead of the more rigorous learning one goes through when trying things themselves.
If you were to ask me what textbooks or lectures I recommend, I think that's a more personal question than many here might guess. What topics are you most interested in? Are you really just solely interested in a solid background? How patient are you when doing problems?
Regardless, I'll give my two cents for textbooks anyway. In no particular order:
1) Griffiths E&M: https://www.amazon.com/Introduction-Electrodynamics-David-J-...
2) Axler, Linear Algebra Done Right: https://www.amazon.com/Linear-Algebra-Right-Undergraduate-Ma...
Get dozens of book recommendations delivered straight to your inbox every Thursday.