Identification of the pregnant woman prone to develop pre-eclampsia later during her course of the pregnancy is a clinical challenge in clinical obstetrics. A new and non-invasive approach to detect abnormalities in pre-eclamptic women that differentiate from women with uneventful pregnancies is presented here. We applied non-linear and fractal features for classifying the dynamical complexity of the heart rate (HR) patterns corresponding to seven normal subjects and eight pre-eclamptic patients. Significant differences in the estimated largest Lyapunov exponent and in the correlation dimension between normotensive women and those with pre-eclampsia were found, suggesting they may have potential as new markers for pre-eclampsia. HR patterns in healthy and pre-eclamptic pregnancies correspond to complex non-linear dynamics, which could arise from the contribution of stochastic and chaotic components. HR of pre-eclamptic patients also revealed a more regular dynamic behavior than those belonging to normal pregnancies, corroborating the general observation that diseased states may be associated with regular HR patterns.