conference - How does the acceptance rate affect papers decisions?

Some conferences do maintain a percentage for the number of accepted papers to the total number of submissions.
Does this mean some good papers will be inevitably rejected just to maintain the acceptance rate? How conference chairs deal with the acceptance rate? or it is not related at all to the acceptance/rejection process.

Answer

Yes.

But it would be more accurate to say that conferences have a fixed budget of papers that they can accept, due to scheduling constraints.

At most computer science conferences, every accepted paper is presented in a 20-minute talk; for a three-day conference with no parallel sessions, this practice imposes an upper bound of about 50 accepted papers. Of course larger conferences have parallel sessions, but program committees generally do not have complete freedom to add another parallel track, partly because of space constraints at the conference venue (which is planned long before the submission deadline), and partly because major changes to the conference organization usually require input from the community.

So inevitably, if a conference attracts a large number of strong submissions, it must reject some of them. This is generally considered better than the alternative, which is that the conference must necessarily accept some bad papers.

This answer is specific to computer science.