Early married on the child-girl is defined as a marriage of a girl before the age of 18. Early married has direct effects to girl education, health, and psychological so that increase risk of the maternal mortality and child mortality cases. The maternal mortality and child mortality cases can also be an indicator of the welfare level of a region. If the maternal mortality and the infant maternal cases are the high, then the quality of health services in the area is bad and conversely. In this paper, we discuss about effect of early marriage on the child-girl to maternal and infant mortality in East Java Indonesia using bi-response Poisson regression model based on local linear. The best model is determined based on Cross Validation (CV). The bi-response Poisson regression model with the best local linear estimator is obtained from optimal bandwidth is 0.67. The comparison of the deviance value between parametric and nonparametric regression model in this case showed the bi-response Poisson regression using nonparametric regression model approach based on local linear estimator is better than the bi-response Poisson regression using parametric regression approach. The analysis showed the early marriage on the child-girl had the highest influence on the number of maternal and infant mortality in East Java.