Why sprint goal is to generate one potential deliverable increment but we always can't finished UAT in UAT env or some team just push the code to SIT

It sounds as though the team are not completing increments of work to the standard necessary for release. How does that compare with your reading of the Scrum Guide, and what it means for work to be "Done"?

What it means for work to be "done" is defined in the team's Definition of Done. This could mean "available for UAT" or "passed UAT", that's up for the team to decide. I typically encourage the team's Definition of Done to only include things that are within the control of the team. If the team, as part of the process, needs to hand work off to someone else, whatever state it must be in prior to the handoff is "done".

Since the Product Owner is a member of the team, it seems reasonable that "passing PO's testing" could be a reasonable part of the team's Definition of Done. If it is and that isn't happening, it could be something to talk about at a Sprint Retrospective.

I will point out that UAT stands for "user acceptance testing". If the Product Owner is performing the testing, it can only be considered a UAT if the Product Owner is a user. Even if the Product Owner is a user or potential user of the system, they may not be representative of all users to adequately perform a UAT.

Thanks all, We need have a confirmation on DOD that we did not do it before.