为掌握南京市住宅地价变动状况，以南京市2004年~2018年住宅用地季度地价指数为研究对象，采用分形理论和重标极差分析法(R/S)进行分析，研究结果表明：1) 南京市住宅地价指数具有明显的分形特征，赫斯特指数(Hurst)为0.7727，时间序列具有较强的正持续性及持久性；2) 南京市住宅地价指数的平均循环周期长度为15个季度，即3.75年；3) 通过累加和变换构建的分形模型可对地价指数进行短期预测，预测结果较为准确，但对地价指数的长期预测误差较大。上述研究，揭示住宅地价的分形特征，科学地掌握了住宅地价波动特点，以期为国土资源管理部门提供决策依据，为社会公众和投资者提供信息参考。
index of residential land in Nanjing from 2004 to 2018 as the research object, the fractal theory and heavy range analysis method (R/S) were used for analysis. The results show that: 1) The land price index in Nanjing has obvious fractal characteristics, Hurst index = 0.7727, and the time series has long-term memory and strong persistence. 2) The average cycle length of the land price index is 15 quarters, or 3.75 years; 3) The fractal model established by accumulation and transformation can predict the land price index in a short-term, and the prediction result is ideal. However, the long-term prediction error of the land price index is more obvious. The above research reveals the fractal nature of residential land price and scientifically grasps the fluctuation law of residential land price. It is mainly for the land resources managers to provide decision-making basis and provide information for the public and investors.