土体龟裂是一种常见的自然现象，会对土体的工程性质产生显著的负面影响，是许多工程地质和环境地质问题的直接诱因，要研究龟裂，首先需要解决研究方法问题。基于国内外近些年围绕龟裂所开展的工作，着重对研究方法进行了系统的归纳和总结，对比分析了各种方法的优缺点，得到如下认识：（1）土体龟裂研究方法总体分为试验研究和数值模拟研究两大类，试验研究又可分为室内试验和现场试验，目前该课题主要以室内试验为主；（2）室内试验常用的试样主要有泥浆样和压实样两种，且前者更受学界欢迎；（3）土体龟裂试验主要涉及两方面的关键仪器设备，一方面是用于控制试验的环境条件，代表性的为环境箱，另一方面是用于监测龟裂发育过程的相关参数，尤其是裂隙几何形态参数，代表性的有图像采集装置、激光扫描仪、CT机、ERT和分布式光纤等；（4）在试验方法方面，多采用控制变量法研究一个或几个参数对土体龟裂形成及发育规律的影响，包括试样初始状态、尺寸大小、厚度、土质成分、环境温度、相对湿度、试样/容器接触条件和干湿循环次数等；（5）采用数字图像处理技术能快速提取龟裂网络的几何形态结构参数，较其他技术更具前景；（6）现场试验可以最大限度的反映真实裂隙的形成和发育规律，但由于这类试验费时费力投入大等客观原因，报道相对较少；（7）数值模拟是研究土体龟裂的重要手段之一，但由于现阶段关于龟裂发育机制没有统一认识，相关研究还处于起步阶段。最后，针对当前的研究不足提出了今后土体龟裂课题的研究重点。

Soil desiccation cracking is a common natural phenomenon, which has significant negative impacts on engineering properties of soil, and is the direct cause of many engineering geology and environmental geology problems. It is a key to choose appropriate methods for studying soil desiccation cracking behavior. Based on the research work conducted on soil desiccation cracking behavior in recent years, the applied research methods are reviewed and summarized systematically. Meanwhile, the advantages and disadvantages of each method are compared and analyzed; and then the following understandings are obtained: (1) There are generally two categories of research methods for soil desiccation cracking: experimental study and numerical simulation. The experimental study contains two aspects: laboratory test and field test. At present, the former one is more popular. (2) There are commonly two types of specimens are used for laboratory test: slurry specimen and compacted specimen, and the former one is also more popular. (3) There are two types of equipment are usually applied in soil desiccation cracking test: one is used to control the environmental conditions like environmental chamber; and another is used to monitor the development of desiccation cracking and the geometrical characteristics of crack network, including crack image capture device, laser scanner, CT machine, ERT, distributed optical fiber and so on. (4) During the test, factors (i.e. specimen initial condition, size, thickness, mineral composition, environment temperature, specimen-container contact condition and wetting-drying cycles) affecting crack initiation and propagation are well studied by control variate method. (5) Digital image processing technique is efficient to quantify the geometrical and morphological characteristics, showing more promising as compared with other techniques. (6) The field test can reflect the real law of the formation and development of cracking; however, very few information can be found in literature because such field test is usually time and cost consuming. (7) Numerical simulation is one of the important methods to study soil desiccation cracking behavior. Because at this stage there is no uniform understanding of the mechanism on cracking, the related research on this topic is still in its infancy. Based on the current research, this paper puts forward the research focus of this topic in the future.