If your PCR is amplifying or looking those genes, which are solely on Mitochondrial DNA so of course that PCR products are from the mt DNA, and its clear that mt DNA encodes around 14 genes, all are well known and most of them are electron transport chain enzymes except Humanin. So its quite simple if you are targeting a gene or sequence which in real not located on mt DNA so how your specific primers will bind to mt DNA in your sample or lysate whatsoever you prepared for the PCR. These PCR products will definitely be from your nDNA.
I'm a little unclear on what you mean. Do you mean you are looking for PCR primers that can specifically amplify mtDNA vs nuclear DNA? Or that you have a PCR product, and you're not sure whether it has arisen from nuclear or mitochondrial genomes? Or something else?
There are obviously many possible sets of PCR primers that could amplify specifically the nuclear and/or the mitochondrial DNA. In terms of the DNA itself: chemically, the mt and nuclear DNA are identical (phosphate, ribose, ATGC etc.); though some features may vary when averaged across the genome - methylation, G/C content, for example - I'm not sure you could say definitely that any particular region was definitely derived from the nucleus or the mitochondria without analyzing the sequence.
One possible thing to keep in mind if you're analysing the DNA without PCR amplification is that the mtDNA is a circular genome rather than linear like the nuclear, which may help you with resolving/degrading one set specifically; however, this may be complicated by the huge excess of nuclear DNA compared to mitochondrial DNA. If you can supply a little more information about your goals/experiments, I may be able to help a little more.
You can check more informations in the following publication: A quick, direct method that can differentiate expressed mitochondrial genes from their nuclear pseudogenes.
Collura RV1, Auerbach MR, Stewart CB. Curr Biol. 1996 Oct 1;6(10):1337-9.
Unfortunately, running on a gel will only work a) on the non-PCR-amplified, nuclear/mitochondrial genomic DNA, and b) if all the fragmented nuclear DNA is of identical size to the mtDNA (unlikely). If, as is more likely, fragmenting your nuclear genomic DNA results in a smear, it will be difficult/impossible to distinguish a band of circular mtDNA from smaller, linear nuclear DNA fragments that run at the same size.
Also, just to add to Jawwad's answer: he's definitely right that there are only a small number of proteing-coding genes in the mtDNA genome (though the exact number varies species to species - 14 is specific to humans and a few others); however, if you're looking for a target for RT-PCR or similar, I'd actually suggest using one of the two mitochondrial ribosomal RNAs, which are also encoded on the mtDNA (along with 20 or so mt-tRNAs); we've done some basic RNA-Seq that suggests that these transcripts are significantly more abundant than the mitochondrially encoded mRNAs.