Background Due to the controversy of the oral glucose tolerance test (OGTT), diagnostic criteria for gestational diabetes mellitus (GDM) in the world and researches on GDM remain undeveloped in China. American Diabetes Association recently recommended the clinicians to diagnose GDM by OGTT results without the third-hour glucose value. This new criteria has not been used in China. Research on the value and sensitivity of the criteria in detecting GDM is rare. The aim of our study is to analyze the characteristics of OGTT in Chinese women with GDM or gestational impaired glucose tolerance (GIGT) and to evaluate the effect of omission of the third-hour plasma glucose (PG) level in OGTT on the sensitivity of diagnosing GDM and GIGT, and the relationship between PG values of 50 g GCT or OGTT and insulin therapy. Methods A retrospective analysis was performed on medical records of 647 cases with GDM from January 1, 1989 to December 31, 2002, and 233 with GIGT. Among 647 cases of GDM, 535 cases were diagnosed by 75 g OGTT. All OGTT results including 535 cases of GDM and 233 patients with GIGT were evaluated. Results There were 112 cases of GDM diagnosed by elevated fasting PG (FPG) without OGTT performed. Of 535 cases of GDM diagnosed by OGTT, 49.2% (263/535) women had FPG value ≥5.8 mmol/L; 90.1% (482/535) women with 1-hour PG values ≥10.6 mmol/L; 64.7% (359/535) with 2-hour PG levels ≥9.2 mmol/L. There were only 114 cases (21.3%) with abnormal 3-hour PG levels among 535 women with OGTT. Among those with abnormal 3-hour PG level, 49.1% (56/114) had abnormal glucose values in the other three points of OGTT, and 34.2% (39/114) with two other abnormal values of OGTT. Our study showed that omission of the 3-hour PG of OGTT only missed 19 cases of GDM and they would be diagnosed as GIGT. Among the 233 women with GIGT, only 4 cases had abnormal 3-hour PG. So, omission of the third-hour glucose value of OGTT only resulted in failure to diagnose 3.6% (19/535) women with GDM diagnosed by OGTT, which means 2.9% (19/647) of all the GDM and 1.7% (4/233) of GIGT in Chinese women. PG levels ≥11.2 mmol/L following 50 g GCT was highly associated with GDM necessitating insulin therapy (75.4%). An elevated FPG level was also associated with insulin therapy (59.7%). Conclusions Omission of the third-hour glucose tolerance test value still yield a higher sensitivity in diagnosing GDM and GIGT. In Chinese women, it is practicable to omit third-hour post-glucose ingestion value of the OGTT in Chinese women. PG levels ≥11.2 mmol/L following 50 g GCT mostly indicates that the requirement of insulin therapy.