I have presented a paper in a conference as a short paper, now I want to submit an extended version of it to a journal without cite the previous paper, this extended version is a completly different version but is the same idea as a short paper. Is it possible to do that ?
If the paper is considered a 'paper' in your field and you are submitting to a journal which requires submission to be new and unpublished, then this is not appropriate. (I am assuming that the two papers are indeed as connected as implied (a short paper, and an extended paper using the same work, discussion, conclusions and adding perhaps more detail but otherwise the same work).
The one scenario that I can imagine in my field (physics) that this might be appropriate is if the conference 'paper' is not a paper, just an extended abstract that is done for the program booklet. If it isn't even something I am going to put down on my CV as a publication, then I don't consider it a 'paper'.
As a referee, when I know a similar paper has been published previously by a group in a conference proceedings or if I check up on their webpage and see that a very similar paper has appeared earlier, it sets up a red flag, particularly if the group did not reference it (hid its presence). It implies to me that the authors are hiding something for a reason.
If the 'new' paper is similar to the un-referenced conference proceeding, but has nothing new in it, then I conclude that the authors deliberately hid the conference paper. My journals require that a paper be 'new' and 'original' and previously unpublished. Writing a paper with different words but the same data, conclusions, main points and discussions with more detail is not new and I would both reject it and point out to the authors and editor the problem. I feel sorry for the authors if they had transparently referenced the prior work (I do also check their work they reference), but not at all sorry for them if they hid it.
If the submitted work has additional data, or a focus on a different aspect of the problem or other features that make it 'new', then I will referee it on its merits and in the review mention the previous work ad suggest they add it to the references. But depending on the relevance of the prior un-referenced paper, it can make me wonder whether they deliberately chose to hide it thinking it would hinder the publication of their new paper. That makes me in future remember the group as potentially suspect.