The great variety of possible configurations of Biological Life Support Systems (BLSS) and regimes for their operation necessitates a “top-down” approach to BLSS development and design. This top-down approach cannot be effectively performed without the criterion of LSS efficiency. In the paper, different criteria for comparing LSS and selecting the most appropriate one are discussed. The most general criterion of success of a space mission inevitably includes reliability or crew safety. Here, the necessity of using the criterion of “integral reliability” of a space mission as a whole is discussed. This criterion incorporates three main indices: reliability, mass, and quality of life. Possible ways of converting mass and quality of life into reliability are considered via examples of LSS launch missions and lunar base scenarios. The unique characteristic of the components of BLSS compared with physicochemical LSS is their ability to self-restore. Accounting for this property may enable ultra-reliable BLSS to be established. The problems hindering the development of technology for building reliable LSS are also considered. Among these, the problems of measuring the reliability of experimental BLSS prototypes and outlining the permissible range of their operation are of key significance.