Needle injection and needle-free injection were proven effective in improving glycated hemoglobin (HbA1c) in type 2 diabetes mellitus (T2DM) patients. However, it is unclear if needle-free and needle injections of insulin during intensive insulin therapy in hospitalized patients provide similar efficacy and safety benefits. A self-controlled cross-over study was conducted on 62 patients with T2DM who received intensive long-acting and short-acting insulin injections with or without needles. The 7-point blood glucose test was performed on the 6th day after insulin administration and the injection method switched on the 7th day of hospitalization. The difference was compared in 7-point blood glucose levels. The blood glucose levels at fasting (mean difference=-1.09 ± 2.38mmol/L, 95% CI, -1.69 to -0.48, p=0.0007) and post-breakfast (-1.14 ± 3.02mmol/L, 95%CI, -1.91 to -0.37, p=0.004) were better when patients were receiving needle-free injections compared to when receiving a needle injection. Indeed, daily blood glucose fluctuation, which presented as the area under the curve of glycemia, was decreased in needle-free injection periods (-0.3.48 ± 9.64, 95%CI, -5.95 to -1.01, p=0.0065). There was no significant difference in the dose of long-acting insulin between the two injection methods (-0.32 ± 2.69, 95%CI, -0.99 to 0.37, p>0.05). The dose of fast-acting insulin during the needle-free period was lower than that of when patients received needle injections (-1.66 ± 6.45, 95%CI, -3.29 to -0.025, p<0.05). There was no significant difference in satisfaction between the two regimens (-0.59 ± 1.55,95%CI, -0.938 to 0.509, p=0.557), but there was a significant difference in pain experience, favoring needle-free injections (p < 0.001). Glycemia was better controlled by needle-free insulin injections in hospitalized T2DM patients subjected to intensive glycemic control. These patients also experienced less pain than when insulin was injected with a needle.