An increasing presence of instream structures such as weirs, dams, culverts and reservoirs degrades habitats, fragments rivers and blocks fish movements worldwide. Longitudinal river movements are fundamental for many fish species and the most widespread solution to restore longitudinal connectivity is the implementation of different fish passage solutions. Fishway functionality, however, is highly variable. To design a functional fishway, several aspects of the fish's interaction with its environment need to be taken into consideration. Artificial light at night (ALAN) can affect a range of different behaviours in fish, from activity and movement to feeding and predator–prey relationships. In a fish passage setting, fish are exposed to artificial light at night (ALAN) in the form of light pollution, but, sometimes, also as part of the fish passage solution. Although likely highly species specific, the effect of artificial light on fish passage behaviour has been little explored. Here we study the passage behaviour of two small‐sized fish species, European gudgeon (Gobio gobio) and Italian riffle dace (Telestes muticellus), over a scaled deep side notch weir in a hydraulic flume in three different light conditions: daylight, darkness and ALAN. Although both species passed the obstacle at high efficiencies under all light conditions, their passage behaviours were influenced by light, particularly at the higher levels. While ALAN reduced passage success and resulted in delayed passage for gudgeon, riffle dace passed at higher rates under the artificial light compared to night treatment. Both results indicate a risk of negative effects from ALAN on passage performance at real fishways—or movement rates in lit areas of natural streams—for both species. Independent of light conditions, individuals of both species also passed faster after repeated trials, demonstrating learning in a fish passage context.