Predicting the ecological impacts of invasive fishes is crucial for understanding their risks in the environment. Experimental approaches that provide empirical data on invasion impacts provide both theoretical insights and data of high management utility for prioritising high-risk invaders. Here, a synthesis of some experimental approaches that predict invasion impacts of non-native fish is presented, where the focus is on impacts relating to the trophic impacts of the invader on either native trophically analogous fishes or prey populations. Experiments in tank aquaria are advantageous in providing homogenous and controlled conditions that also enable high replication. Competition-based studies can determine invasion impacts via exposure of fish to fixed food rations, although these experiments tend to produce results in intensely competitive conditions that might not occur in the wild. Comparative functional response experiments in tank aquaria have been successfully applied to determining the relative impacts of invaders compared with native species on prey populations, and have revealed high resource efficiency in globally invasive piscivorous fishes. Experiments completed in mesocosms and small ponds have the advantage of providing replicated systems in semi-natural conditions. They have been successful in revealing that rather than competing with native fishes, many invasive fishes show strong patterns of trophic niche partitioning. These patterns are also evident in these fishes when in natural invaded communities. Thus, whilst it is often suggested invasion impacts from non-native fishes result from inter-specific competition, evidence for this from experimental and field-based predictions is equivocal.