In this article, we propose a neurophenomenological account of what moods are, and how they work. We draw upon phenomenology to show how mood attunes a person to a space of significant possibilities. Mood structures a person’s lived experience by fixing the kinds of significance the world can have for them in a given situation. We employ Karl Friston’s free-energy principle to show how this phenomenological concept of mood can be smoothly integrated with cognitive neuroscience. We will argue that mood is a consequence of acting in the world with the aim of minimizing expected free energy—a measure of uncertainty about the future consequences of actions. Moods summarize how the organism is faring overall in its predictive engagements, tuning the organism’s expectations about how it is likely to fare in the future. Agents that act to minimize expected free energy will have a feeling of how well or badly they are doing at maintaining grip on the multiple possibilities that matter to them. They will have what we will call a ‘feeling of grip’ that structures the possibilities they are ready to engage with over long time-scales, just as moods do.