Mg2+ concentration, higher amount of EDTA, too much template conc or improper denaturation, pH of the reaction buffer, salt impurities in water in use these are the major inhibitors other than the storage condition of the enzyme itself.
Are you getting shorter reads in sequencing? Smaller fragments than expected? Hairpins in your template may cause the polymerase to fall off prematurely. Try adding a denaturing agent such as Formamide, DMSO or Betaine if the GC content is high, or increase the annealing temperature as high as possible to help keep the DNA single stranded. Using a different polymerase with better processivity (such as Phusion) will also help.
There are a lot of factors that can cause premature termination in a PCR. To what others have already said, I would add sequence complexity, such as high degrees of secondary structure.
If you don't find an easy fix to your problem in going over your notes, then I'd suggest optimizing your PCR. Load multiples of each sample in different PCR conditions. Using your past PCR as a baseline for measurements, vary each separate PCR condition by only one change in protocol (annealing temp, type of polymerase, etc). Hopefully, this will tell you what PCR conditions are best for your particular reaction.
Regarding optimal annealing temperature, this can often be overcome by using a touchdown PCR protocol.
Similar to secondary structure, regions containing poly Gs can be as problematic. I would agree with Linda that PCR additives like DMSO or Betaine often work. We would sometimes use Q-solution, a reagent included in the PCR kits from Qiagen. Although, this may only be Betaine. If all else fails, it is worth trying.