A tree line, as the altitudinal or latitudinal limit of erect trees, is considered as a sensitive ecological indicator of climate change, and becomes one of the hot issues in the studies of global change ecology. During the last decade, rapid progress has been made in tree line studies. However, field survey and research methods may vary significantly among tree line literatures, limiting the evaluation of mountainous forest response and feedback to climate change at regional or global scale. Herein, we reviewed the research progress regarding the field survey and research methods on tree lines, evaluated the advantages and disadvantages of each method, and pointed out the current research frontiers. Field survey and research methods in tree line literatures mainly include: Repeat landscape photography, remote sensing image analysis, land line transect method, circular sampling plot, rectangular or square sampling plot, spatial point pattern analysis, tree line dynamic model, controlled experiment, and transplant experiment. Repeat landscape photography and remote sensing image analysis can provide an intuitive reference for treeline dyanmics, but some uncertainties remain. Compared with selective sampling approach (e.g., line transect method, circular sampling plot and square small plot), sample-total method (rectangular large plots including the whole tree line ecotone, i.e., encompassing the current timberline and the tree line) provides more robust results regarding tree line structure and shifts. Spatial point pattern analysis has been used to establish the linkage between the ecological patterns and processes of the tree line ecotone. Tree line dynamic models can be used to reveal temporal patterns of position and structure of tree line ecotones and their driving mechanisms. Controlled or transplant experiment has advantages in exploring the critical drivers of tree line dynamics. In future studies, sample-total method and its protocol are recommended when exploring variations in structure and position of tree lines; regular monitoring of fixed large tree line plot is worth carrying out; controlled or transplant experiment can be set up at diverse tree lines across a regional scale; researchers should attempt to develop new tree line dynamic models with good transplantation capability.