Immersion in forest environments was shown to produce beneficial effects to human health, in particular psychophysical relaxation, leading to its growing recognition as a form of integrative medicine. However, limited evidence exists about the statistical significance of the effects and their association with external and environmental variables and personal characteristics. This experimental study aimed to substantiate the very concept of forest therapy by means of the analysis of the significance of its effects on the mood states of anxiety, depression, anger and confusion. Seven forest therapy sessions were performed in remote areas and a control one in an urban park, with participants allowed to attend only one session, resulting in 162 psychological self-assessment questionnaires administered before and after each session. Meteorological comfort, the concentration of volatile organic compounds in the forest atmosphere and environmental coherence were identified as likely important external and environmental variables. Under certain conditions, forest therapy sessions performed in remote sites were shown to outperform the control session, at least for anxiety, anger and confusion. A quantitative analysis of the association of the outcomes with personal sociodemographic characteristics revealed that only sporting habits and age were significantly associated with the outcomes for certain psychological domains.