- I submitted my conference paper to IEEE ICC. It took nearly 4 months in decision.
- During waiting time, I also submitted its 2 extended versions to 2 different journals.
- The conference paper met acceptance and I uploaded camera ready version, but the conference is yet to be held and therefore the paper is still unpublished.
- Meanwhile, my journal papers also met acceptance and appeared online. Now, when I check plagiarism of my unpublished conference paper, it is 37%, higher than allowed. Previously it was 8%.
- My question is will IEEE be able to publish or reject my paper at this stage?
While I think your intentions were good, you seem to have made a bit of a mess of it. I can't predict what a journal or conference would do about retracting acceptance, but if there is apparent self-plagiarism, they might decide to do that.
It is hard to say exactly what you should do now to protect yourself and also get everything published. What you should have done earlier, however, is to cite the earlier papers, even if not yet published, in any later ones. That citation makes any self plagiarism charge impossible.
I'll not again that the purpose of avoiding self plagiarism is to allow readers of your work to trace what they read back to the original source of the ideas, including the complete context of the earlier papers. To cite an unpublished work that has been submitted, just mark the citation as "to appear".
I don't know if it is possible to work with the various editors at this point to provide updated versions that correct the issue. If they allow that, then you should come out ok and also get the various papers published as usual. You should probably alert them immediately of the issue, providing an explanation (and maybe an apology for your misunderstanding), as well as corrected versions.
It would probably be good if you contact them before they make any decisions, but I can't say for sure. Their automated systems have already flagged your work.
For completeness, note that if you have no intention of publishing some work, then you don't need to cite it. Treat it as nothing more than an early draft of the later work.