Endobronchial ultrasonography with a guide sheath (EBUS-GS) has recently been used for improved diagnostic yields for peripheral pulmonary lesions. This study retrospectively evaluated the factors related to the diagnostic yield of EBUS-GS for peripheral lung cancer. The medical records of 76 patients who had been diagnosed with lung cancer and had undergone bronchoscopy with EBUS-GS in our hospital between August 2014 and September 2015 were reviewed. The total diagnostic ratio of peripheral lung cancer was 71.1%. The following factors of the diagnostic yield were evaluated: location of pulmonary lesion; size; feature; bronchus sign; location of EBUS probe; EBUS detection; number of biopsies performed; procedure time; use of virtual bronchoscopic navigation; use of EBUS-guided transbronchial needle aspiration with EBUS-GS; CT slice thickness; operator's years of medical experience; and specialized training in bronchoscopy at the National Cancer Center. In all cases, lesion size ≧ 20 mm (80.8% vs. 50.0%, P = 0.006), EBUS probe location "within" (90.0% vs. 50.0%, P < 0.001), EBUS detection (80.7% vs. 28.6%, P < 0.001), number of biopsies ≧ 5 (78.0% vs. 47.1%, P = 0.013), and bronchoscopy training (81.6% vs. 60.5%, P = 0.043) significantly contributed to an increase in the diagnostic yield. Following a multivariate analysis, EBUS probe location "within" was found to be the most significant factor affecting the diagnostic yield (odds ratio 14.10, 95% CI 3.53-56.60, P < 0.001), and bronchoscopy training was the second most significant factor (odds ratio 6.93, 95% CI 1.86-25.80, P = 0.004). EBUS probe location "within" and bronchoscopy training are the most important factors for improved diagnostic yield by bronchoscopy with EBUS-GS for peripheral lung cancer.