Dietary omega-3 polyunsaturated fatty acids (ω-3 PUFAs) can be classified into animal- and plant-derived ω-3 PUFAs. Patients with type 2 diabetes (T2DM) are frequently accompanied by dyslipidemia, which is closely related to the high-density lipoprotein (HDL-C) subfractions change. This study aimed to determine the effects of different sources ω-3 PUFAs on glucolipid metabolism and lipoprotein subfractions in T2DM with dyslipidemia. Ninety T2DM patients with dyslipidemia were randomly assigned to take 3 g/day fish oil (FO, containing eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA)), 3 g/day perilla oil (PO, containing α-linolenic acid (ALA)), or 3 g/day blend oil (BO, containing EPA, DHA and ALA) for 3 months. 90 patients completed the intervention. There was a significant reduction of glycated hemoglobin (HbA1c) in all the groups. The triglycerides (TG) in the FO group were significantly different with a group × time interaction (P = 0.043), which was higher compared with the other two groups. The serum small HDL-C subfractions in the PO group was higher and the serum large HDL-C subfractions in the PO group was lower than those in the BO and FO groups. Plant-derived ω-3 PUFAs are more effective at controlling blood glucose than animal-derived ω-3 PUFAs. However, animal-derived ω-3 PUFAs have a significant lowering effect on TG compared with plant-derived ω-3 PUFAs. Particularly, large HDL-C subfractions after animal-derived ω-3 PUFAs intake were higher than plant-derived ω-3 PUFAs intake; while small HDL-C subfractions were lower. Both the animal- and plant-derived ω-3 PUFAs have practical value in improving glucose and lipids metabolism in T2DM patients with dyslipidemia.