Mainly Due to Heterogenity/Polydispersity of your sample if your system suitability is passed. 2. might be problem in acrylamide quality 3. Artifect : your plates are not properly cleaned 4. Overstaining or not properly optimized staining procedure.
Was going through some old discussions and found this discussion thread. Sorry it may not be relevant anymore but just thought will be on record for others.
In addition to the reasons listed by Yamak, from my experience smearing can also result from incomplete dissolution of the sample in the PAGE loading buffer. You could increase the % of SDS in the loading buffer to 4% or more. Then allow sufficient time for dissolution. Alternatively, you could dissolve the protein matrix in 10% SDS first and use it for making dilution in PAGE buffer.
Lastly make sure that your sample has adequate reducing agent. If your sample (in PAGE loading buffer) is too old you could add some DTT (usually 100 mM, dependent on your protein conc) and re-check again.