How to improve one's programming this language from beginner to higher?

Here's what you need to do, download programs and then disect them and experiment with the commands and things you find. That's how I did it.

Back in the Petit Computer days I would find programs that I liked and I would dissect them to see how they worked. I would learn how a line of code works by disabling it with a apostrophe. That would turn that line of code into a comment that isn't read in the program. I would also spend time reading the instruction list and playing with the different commands to see what they did. I posted some examples you can look at in your "How to make it so that the sprite's collision stops the player/player's sprite from going through the sprite from ALL sides?" post. Good luck! :)

Trial and error is some way I did it, aside from looking at how other people did stuff. I always thought "if something doesn't work, try it another way".

Also statrting a small project and working your way up to make it a more complex project helps

I tried dissecting projects and got very limited results. Trial and error only works to a certain extent for me(from my other learnings of programming languages). Trying out different commands helps a bit. Making a small project into a large project gets me limited results. I wish I could do this with a well-written book (like a book on a college level written like a Pearson publisher book) or through school. I am good at school learnings.

I really recommend just picking a tutorial for something simple you want to make and trying to translate it into SB code. There are thousands of tutorials on the internet about making games. There's not always a right answer, and you're going to have to apply critical thinking eventually. Here's a block programming puzzle for Flappy Bird if you're not comfortable reading another language: https://studio.code.org/flappy/1

I think that each language requires it's own study so studying other languages and general programming techniques won't help. But I'm not sure if I'm right saying that.

I think that each language requires it's own study so studying other languages and general programming techniques won't help. But I'm not sure if I'm right saying that.

You aren't ;) The general premise of programming is manipulating the shoeboxes that represent your memory. These shoeboxes are the same for every language. And Flow control such as IFs and FORs are the bread and butter of almost any language. IFs (branches) are even built into the processor. A large portion of the problems you are solving when you programming are actually logic problems. That's where the term "algorithm" comes in. Its logical set of steps or "programming" that isn't in any computer language. You could take any algorithm for AI pathfinding and rewrite it in any language, and it will behave the exact same - some languages may be faster for the way data is stored, but the concepts translate directly across. That's why I suggested you take a look at the block programming example. It conveys the logical steps you would have to take to achieve the end goal, without ever making you write a specific piece of code.

How about getting a notebook, to write what you learn or want to learn in? It's working for me so far and if I need to know more about a specific command, I just go to the chat : smilebasicsource.com/chat

I think that each language requires it's own study so studying other languages and general programming techniques won't help. But I'm not sure if I'm right saying that.

Yes, you're right, each language requires its own study. Especially when it comes to different 'paradigms', such as object-oriented, functional, actors, and the like. But you're also wrong - generally, the more programming languages you know, and the more techniques you know, the better a programmer you will be in each of those languages. Even for switching to another language within one 'paradigm', things like having a data structure such as a linked list or associative array as a primitive data type will help you appreciate data structures better, help you devise more potential solutions to a problem so you have a wider range of choices from which to pick the best solution.