i think if we reduce the quantity of primers and sample DNA it can reduce the chances of non specific bands. how much quantity of primers are you using?

Adjusting the temperature properly will also help besides reducing the quantity of primer and DNA. While temperature standardization try to find out highest temperature with clear banding pattern as lower temperature encourage mispriming.

hard to answer without a picture but some aspects are..design the primers using online software like primer3. Do not use too much dna. Do not use too much magnesium. Run an annealing temperature gradient to reduce the intensity of non specific bands. Add DMSO in increments up to 6% to increase specificity. Consider using touchdown pcr where the early cycles use annealing temperatures above where the lower annealing primer can bind to enrich the reaction mixture with specific product. For instance if your primers anneal at 54 and 58 start annealing at 62 and work downwards to 52 then cycle at 52 annealing for the remainder of the cycling. Use a hot start enzyme often helps.

Thank you so much Muhammad Nauman, I use 1 ul of 10 pmol of each Reverse and Forward primers. and also Thank you so much Sundeep Jain and Paul Rutland for your valuable guidance.

You can also use a (basic) oligonucleotide analyser for your primer sequences like this: https://www.idtdna.com/calc/analyzer to assess how likely it is that your primers will form hairpins, homo and hetero-dimers.

You need to select primers with proper melting temperatures.

All the previous advices are actually the most valuable.

Anyway, sometimes we are forced to use "bad" primers, so I add something to do in 'desperate' cases.

1) keep the template quantities as suggested by the polymerase datasheet, but prepare a series of reactions with different concentrations of the two primers: more than a few times you will find a condition in which non-specific bands are minimized.

2) when non-specific bands are higher than the specific one, try by reducing the annealing and elongation times, for disadvantaging the former and favor the latter.