Abstract Background Caesarean section (CS) is a procedure that can be lifesaving for both mother and child. In many developing countries, in particular countries in sub-Saharan Africa and South Asia, access to this lifesaving procedure is low, resulting in high maternal and infant mortality rates. In many high- and middle-income countries the procedure is extensively performed, and the reasons not always medical. The WHO recommends a CS rate of 10-15%. In Nepal the rate is 5%, according to WHOs latest report from 2014. The aim of this study is to investigate the rate and indications for performing CS in Okhaldhunga community hospital (OCH), Nepal. Method The study is a cross-sectional prospective study. All women undergoing CS from the period 07.08.2014 to 15.11.2014 were interviewed regarding socioeconomic background and medical history using a semi-structured questionnaire. Information about the birth and outcome was obtained from the medical record and the birth chart. Results There were 251 births at OCH during the 14 weeks study period. Out of these, 22 women underwent CS and 229 had vaginal delivery, giving a CS rate of 9%. The most frequent main indications for CS were foetal distress (3 women, 14%) and prolonged labour (3 women, 14%). Emergency CS accounted for 2/3 of the sections while elective CS for 1/3. The nulliparas had a lower rate of elective CS (9%) compared to the multiparas (55%). Discussion and conclusion The rate of CS conducted at OCH was 9%, which is a low rate. The CS rate is probably even lower, as 67% of deliveries in Nepal is conducted at home. 2/3 of our data were emergency CS. This is a high number of emergency CS, but it is as we would expect in a rural area with long distances to health facilities. Elective CS was more frequent among the multiparous than the nulliparous. This might be due to indications regarding previous deliveries like previous CS, previous traumatic delivery and disproportion. Our data is too limited to do statistical tests and draw any conclusions, so we will countinue to collect data untill May of 2015, and hopfully publish our full work then.