1. I read a lot of literature. At present, on average, I can only write one article after reading about 20 articles. Writing down common English while reading literature is very helpful for you to write articles. I think a math paper takes about 2-3 days, a short article, and one article a day. Reading documents in the early stage is actually a very efficient thing, which is far easier than looking up information when writing problems.

2. Strictly control the quality of references. The magazines I contribute, including my collaborators, will consider the quality of the articles cited in the references. For example, if you want to send SCI, the references should be above SCI. There is also timeliness, and there must be the latest content. Unless there are special circumstances.

Write directly in English, otherwise you can't guarantee the grammar is completely correct. If the sentence is true, at least write it in English. Then find someone to modify it, preferably an expert in the industry. I found a foreign teacher to graduate with me. Of course, I am the first author.

4. Write more and cast more, as @ Pig said.

You can look at the requirements of the journals you want to contribute, and some journals will stipulate the length. As @ Pig said.

6. If you want to know the speed of periodical publishing, I will look at the time from the "receiving-receiving-sending" in recent issues. If there is a backlog of manuscripts, or there are special reasons (for example, several issues are reserved for some important conferences), information that requires the author to contribute cautiously will generally be published on the first page of the journal.

7. SCI journals of mathematics will definitely require excellent innovation. If it is an imitation of a work or a statement of some ambiguous promotion, it is basically impossible to publish it in SCI.