Objectives This study provides an evolutionary perspective to a classic topic in demography, that is, the discrepancy between reproductive intention and subsequent behavior, in the context of China's two‐child policy. Methods We conduct an event history analysis of longitudinal data from the 2015 and 2018 waves of the Xi'an Fertility Survey (sample size = 321 followed one‐child mothers) to test the hypotheses of how within‐family support/conflict affects women's fertility behavior. Results Only 50% of positive intentions (i.e., intending to have a second child) led to another (live) birth within the 3‐year interval; meanwhile, 15% of uncertain intentions and 5% of negative intentions resulted in a birth. Husband's and the firstborn's emotional support raised the hazard of second childbirth along maternal life course, which cannot be fully mediated by mother's fertility intention and thus, contributed to an intention‐behavior gap. Husband's sibship size had dual effects on female childbearing behavior: A positive indirect effect mediated by fertility intention, but a negative direct effect presumably due to sibling competition for intergenerational support. Finally, after controlling for fertility intention, having a firstborn son was still associated significantly with a lower second‐childbirth hazard, presumably due to son preference as well as concern over parental investment. Conclusions Our study identifies a discrepancy between maternal fertility intention and realized childbearing, which was partly explained by (lack of) support from other (multiple) stakeholders in family reproduction.

After a universal two-child policy was introduced in China in 2016, studies have been published using women’s fertility intentions to forecast future births; however, the recommended algorithms need to be improved. In this study, an algorithm based on the method of limiting factors is developed to retrospectively forecast annual births in Xi’an City in the first three years of policy implementation, i.e. 2016–2018. The 2015 Xi’an Fertility Survey (sample: 560 one-child mothers) showed that 17% of mothers intended to have a second child, 30% were undecided and 53% did not intend to do so at the end of 2015. The low forecast variant based on the updated algorithm indicates that there would be a baby boom in 2016–2018, but the annual births would increase by 13% at most. The forecasting results are basically consistent with the official reports on annual births. This study emphasizes the importance of appropriately adjusting all fertility intentions in birth forecasting, helps to understand women’s fertility behaviour and evaluate the effects of implementing the universal two-child policy, and has important implications for China’s population and family planning work.