High quality forest regeneration is typically performed by a combination of site preparation through mounding and deep planting. These operations are performed either by mounders with subsequent manual planting, or by mechanized tree planting devices that creates mounds and, simultaneously, plant seedlings. These forest regeneration strategies are sustainably beneficial and have high development potential particularly regarding efficiency. Recent development in digitalization, sensing, IoT etc. has enabled new forest regeneration strategies. Hence, the objective of this thesis is therefore to develop a methodology for how automation can be used to improve sustainability in forestry and in forest regeneration operations in particular. The research has followed the Design Research Methodology in which success criteria was identified concurrently with analysis of the as-is situation. Then different sub-solutions forming the automation methodology was prescribed and validated in subsequent descriptive studies. To enable automation, it was found that data needs to be retrieved from surroundings while performing mounding or mechanized planting, whereby analyses of this data need to be used to make improved decisions of the machines’ working procedures. First, properties and characteristics of the surroundings on a clearcut were defined. Then, different ways to retrieve data from the surroundings were tested and evaluated. In addition, several ways of analyzing such data for improved decision making were found and validated. The information gained from data collection of surroundings and subsequent analyses further has potential to be improve activities beyond forest regeneration e.g. smart forwarding, customer adapted assortments, history tracking etc. To use as support for how decision making can be improved from such information, an experimental terrain vehicle platform for researchers and machine developers was developed to enable tests and validation of new solutions with special focus on autonomy and robotics, which requires a variety of data collected from both exteroceptive and proprioceptive sensors. The resulting methodology shows how forest regeneration can be automated in the short-term. Developed methods for data collection and analysis can enhance obstacle avoidance for mounding significantly, and thus contribute to increasing the share of continuously intermittent mounding conducted in Fennoscandia.