this can happen due to many reasons such as contamination of samples and/or primer mix, Pcr specifically annealing temp., specificity of primers and Amount of template too. Therefore, better to tackle every single possible problems one by one so that you can pin point the problem. for contamination, just use extra measure for your working environment and materials and be extra conscious for your steps while working for not to cross contaminate, use every thing new even new primer mix to start with. For annealing use some reference information that are specifically related to your work then go up and down from that and see how it goes. For using literature information of different researchers probably you can find different types of primers so try a couple of them and see which one works better for your samples. For template it depends on the impurity of the extract and concentration just by simply adding more template DNA to PCr will not give you better result because impurity will also increased so that it can affect your result.

In general, PCR needs more detailed observation and repeated trial to conclude that the result is acceptable so even if you get what you expect do not for get to try one or two times again to be conclusive. Reputability of the experiment is the key to final conclusion.

Good luck 

I wonder if you have optimised the PCR condition before you started PCR the real samples, especially MgCl2 concentration & Ta (calculated according to the GC content or suggested Temp by the company you bought the primers nowaday; you then test  +/- 0.5C of Ta ). 25ng or 50ng DNA will be fine 30 -33 cycles. Referring to your gel image, they seem to be all samples without controls. I always use negative and positive control with each set of PCR to ensure that there's no contamination or failed experiment. Hope it helps.