Numerous symptoms in psychiatry are subjective (e.g., sadness, anxiety, craving or fatigue), fluctuate and are environment dependent. Accurate measurement of these phenomena requires repeated measures, and ideally needs to be performed in the patient's natural environment rather than in an artificial laboratory environment or a protected hospital environment. The usual paper and pencil questionnaires do not meet these two conditions for reasons of logistics. A recently developed method, ecological momentary assessment (EMA), made it possible to implement these field assessments via ingenious use of various devices (most frequently an electronic diary) coupling an auditory signal with computerized data capture. The subject carries the device with him/her at all times, and data is recorded in vivo in real time. The programming of repeated measures in the form of a Likert scale or pull-down menu is easily achieved. A recall alarm system can help increase compliance. Compared with classical self-report, EMA improves the validity of the assessment of certain symptoms, which are the main evaluation criteria in clinical trials concerning certain pathologies (e.g., craving and treatment of addiction), where measurement was previously liable to bias. This article sets out to present this method, its advantages and disadvantages, and the interest it presents in psychiatry, in particular via three original applications developed by the authors including: measurement of reaction time without the knowledge of the subject in order to test certain cognitive models; use of a graphic solution for the data recorded for functional analysis of disorders; and the use of data collection via mobile phone and text messages, which also enables therapeutic interventions in real time by text messages, personalized on the basis of the situational data collected (e.g., in the case of craving, the associated mood, solitary or group consumption or concomitant occupations).