China has since the mid-1980s witnessed an increasing trend in sex ratio at birth (SRB) imbalance. While the literature agrees that SRB is caused by parents' need to have a son, it is not clear if an increase in SRB reflects an increase in son preference. This study aims at bridging this gap by developing a model to estimate prenatal son preference based on SRB and total fertility rates (TFR). Data on SRB and TFR from China are applied to the model. The results show that since the mid-1980s, prenatal son preference has fluctuated, and trends in SRB and prenatal son preference by province and by rural, township and urban areas sometimes diverge. In order to interpret these results, changing trends in prenatal son preference is discussed in relation to social change in China. The study draws conclusions at two levels. First of all, theoretically, it concludes that increased SRB does not have to signify increasing prenatal son preference when fertility rates fall. Secondly, it concludes that in China, the process of socioeconomic development has diverging impact on son preference, weakening it in some instances, while reinforcing it in others. Modeling son preference as suggested is useful for posing relevant questions with regard to how and why son preference changes. It also has important implications for where, if and how interventions that address skewed SRB should be planned.