自上世纪70年代实施计划生育政策以来，我国的生育率持续走低，目前已经长期保持在更替水平以下。过低的生育率将带来严重的社会经济问题，如人口红利的消失、劳动力不足、人口老龄化严重等，如何通过生育政策的调整和实施改变生育率持续降低的趋势是人们关心的问题。生育意愿是指个人根据自身条件、家庭状况、社会经济发展情况以及生育政策等因素综合考虑所做出的未来生育决策，包括生育数量、时间等维度。生育意愿一般被认为是实际生育水平的先导指标，研究影响生育意愿的因素可以前瞻的了解到哪些因素会影响实际生育水平，进而为今后的生育政策的制定提供参考和依据。因此分析影响生育意愿的因素非常重要。本文利用中国健康与营养调查2011年的横截面数据，选取52岁以下的已婚女性作为样本，女性总共愿意生育孩子数目是否大于1作为因变量，共选取了有关个人特征、家庭情况、社会因素的26个自变量作为潜在的影响因素进行分析。运用样本数据建立Logistic回归模型，并用Lasso惩罚回归方法估计系数，以达到变量选择的目的，采用交叉验证（Cross Validation）方法选取合适的惩罚系数以得到较优的模型。首先用选取的全部样本建立模型，然后合并某些分类较多的变量的类别再次建模，最后把样本分为城市和农村两个样本分别建模进行比较。本文分析发现，女性所处省市、最高学历、户口类别、工作单位性质、年龄、父母是否健会影响女性的生育意愿，并且影响城市女性和农村女性生育意愿的因素稍有不同。本文最后建议根据不同地域、不同受教育程度的女性实施有差别的具体生育条例，鼓励生育以调整生育率下降的趋势。

Since the late 70's, China’s family planning policy has took effect on the fertility rate, which has been declining and below the replacement level during these years. If the fertility rate remains too low for a long time, it would lead to a series of problems such as the disappearance of ‘demographic dividend’, shortages of workforce, rapid ageing population and so on. China manages to change this ‘low fertility rate’ trend through adjusting family planning policy and ways of implementing. Fertility intention refers to how many and when a couple would like to have children based on the considering of condition of themselves, family and the society. The gender of children is also considered. This is a variable which has a strong impact on the actual fertility rate. Thus studying factors that influence fertility intention is important for analyzing fertility rate. Data used for studying comes from China Health and Nutrition Survey (CHNS), 2011. Married women under 52 was selected. Whether they would like to have more than one child was regarded as dependent variable, and 26 factors involving individual, family and society are included as independent variables. Logistic-Lasso regression was preformed to estimate coefficients, and Cross Validation was used to select best model. Firstly, a whole sample model was built and then I merged some categories to rebuild the model since some variables have too many categories which is not easy to interpret. At last samples from urban and city are used to build models respectively for comparing. After analyzing, I found that provinces, education, urban/rural, company ownership structure, age and alive parents are effecting a woman’s fertility intention. What’s more, the factors is a little different for urban women and rural women. At the end of this article, I suggested that our country should implement different family planning policy based on different province and woman’s education level.