Please show more views of your model. Is there a location with a sharp inside corner like this?
All inside corners near areas of interest should have at least a small radius. For the plastic parts I work with 0.15mm radius is used. Compared to H-elements, P-elements are less forgiving. In H-elements the stress is averaged over the element size which hides the singularity until using smaller and smaller elements. For P-elements the order of the element captures the high stress with larger elements. You can artificially limit the order of the elements near singularities but I prefer a tiny, but realistic radius.
What strain is it reporting vs. what you expect? Are you using a force or an enforced displacement. With a force this same message often comes when the design cannot reach the load applied, sometimes even failing to finish the loadsteps. I also suggest to plot a measure of plastic strain or stress vs loadstep. You can stop the analysis and view intermediate results at each loadstep.
Finally element quality/number will affect the ability to model large strains. A mapped mesh with quads and more layers through the thickness will help capture larger strains without too much element distortion. A tetra mesh with large elements could contribute to the issue.
I try to improve my model, but I obtain again high stresses;
I added a small radius where the displacement occurs and mapped mesh
In this way the values of stress down, but there is again the warning
The stress is much higher than the yield strength of one of the
elasto-plastic materials in your analysis. Please review
your material properties and/or units."