Objective　 To study the clinicopathologic relevance of a thymic epithe lial tumor (TET) grading standard with the WHO classification. Methods　 A gradi ng system for TET was proposed based on the application of WHO histological typi ng of thymic tumors and analyzed in relation to clinical therapy results and fol low-up data of 200 TET cases. Results　 In this series, 8 patients (4.0%) belong ed to type A, 68 (34.0%) were type AB, 17 (8.5%) were type B1, 39 (19.5%) were t ype B2, 27 (13.5%) were type B3 and 36 (18.0%) were type C. The remaining 5 case s were rare thymomas. The overall postoperative survival data showed highly sign ificant differences among the histological subtypes (P<0.001). Typ e A & type AB thymomas showed excellent prognosis, none of these patients died of tu mor; in type B1, only 1 case (5.9%) died at 22 months postoperatively. Types B2, B3 and C thymomas shared the bad, worse and worst prognosis. Ninety-six patien t s (48.0%) were in stage I, 26 (13.0%) in stage II, 65 (32.5%) in stage III and 1 3 (6.5%) in stage IV. Clinical stage is also highly significant in predicting su rvival (P<0.001). It was found that tumor histology could predict surv ival expectancies well in stage I and stage II cases. It was also found that typ e B2, B3 and C thymomas had a statistically significant worse prognosis than typ e A, AB and B1 thymomas (P<0.001). According to the histology, cli nica l data, biological behavior and prognosis, it is proposed that thymomas be divid ed into 4 grades: grade I, II, III and IV. Follow-up is the best strategy for g r ade I & II patients after radical surgery. In this series, the 30 patients (15.0 %) presenting clinical signs of myasthenia gravis were mostly in type B2 and B3 groups (P<0.01). Conclusions　 The WHO classification for TET provides good patho logical definitions and criteria for diagnosis, which can independently predict the invasiveness and prognosis of TET. TET grading is of use in unifying patholo gical and clinical findings, in selection of proper therapy and in predicting pr ognosis.