The place conditioning paradigm is an efficient, widely-used method to study mechanisms that underlie appetitive or aversive learning and memory processes. However, pharmacological agents used to induce conditioned place preference (CPP) or aversion (CPA) can per se interfere with learning and memory processing, hence confounding the results. Therefore, non-pharmacological place conditioning procedures are of high importance. Here, we introduce a novel procedure for induction of CPA in mice, by water flooding. We found that pairing a context with immersion in moderately cold shallow water resulted in aversion and avoidance of that context during a place preference test. Importantly, place aversion emerged only when mice experienced the onset of flood during conditioning training, but not when mice were placed in a compartment pre-filled with water. We also found that warm water was not sufficiently aversive to induce CPA. Moreover, CPA was observed after two or three context-flood pairings but not after one or four pairings, suggesting that moderate conditioning intensity produces optimal CPA expression. Thus, flood-induced CPA is a simple, cheap, and efficient procedure to form and measure place aversion memories in mice, using an ethologically-relevant threat.