Maternal Mortality Rate (MMR) in East Java was still high. in 2013, MMR in sidoarjo district has readed 96.27 per 100,000 live birth. This aim of this study was to the effect of the referral process to maternal mortality in RSUD Sidoarjo. This research was analytic observational with case control design. The Samples of this study were 25 pregnant women who were referred to RSUD Sidoarjo and death. The case controls were 50 pregnant women who were referred to RSUD Sidoarjo who did not experience death. Techniques of data collection using secondary data from the register book maternal and neonatal Emergency (MNE) and medical records and interviews with the mother/family/husband of respondents. The data was analyzed by using univariable, bivariable and multivariable analysis with logistic regression. The results of this study confirmed that the referral process was poor (OR=9,783,95% CI: 2,275 to 42,072, p=0,002) and the complications (OR=0,005,95%CI: 0,001-0,057, p=0,000).thus, the incidence maternal mortality increased. The conclusion of this study is the referral process and the complications to maternal mortality affect the occurrence of maternal mortality. Midwives need to conduct health education should be given to women in their productive age, increase the participation of families, communities and cadres in the process of early detection of complications during pregnancy, childbirth and postpartum, the quality of antenatal care (ANC) and the quality of referrals should be improved by creating a close referral system in a region associated with a high risk pregnant women were detected inventoried and scheduled control/termination and monitored (follow-up) so that high risk always monitored.