Hi @sajjad.sarikhan,
Sorry but I'm not sure I quite understand still what you are trying to do.
When you say "qiime2 database correspondong to my samples", what exactly are you referring to?
If your end goal is to have a separate feature-table, phylogenetic tree, representative-sequences.. etc for each sample instead of the normal combined approach, you can achieve this using a combination of some of the available filtering methods after creating the combined version of said files.
For example, you can use filter-samples to retain only say Sample1 from your combined feature-table. This would leave you with a second feature-table which only has Sample1. Now, if you wanted the representative-sequences of Sample1 alone too, you could use that second feature-table and the filter-seqs action to filter and retain only rep-seqs for that sample from your combined rep-seqs file. Using the newly developed rep-seqs file then you can create a new phylogenetic tree for Sample1 alone using the standard tree-building tools within qiime2 phylogeny.
This of course still doesn't provide you with your original inquiry of separate fasta files per sample after dada2, but perhaps it does get you where you want? And my apologies in advance if I misinterpreted your intentions completely. 
If I'm way off, perhaps an example of what you want may help us get you there.