The influence of fish on the mooring loads of a floating net cage is studied numerically and experimentally. Two experimental series were conducted. One case was model tests with artificial fish. Nine rigid fish models with total volume of 2.5% of the fish cage at rest were placed inside the net cage without touching the net and towed with the net cage. The other case was live fish experiments in waves and current where more than 800 salmons of length 16 cm occupied about 2.5% of the fish cage volume at rest. The flow-displacement effect of a rigid fish in current was simulated by a potential-flow slender-body theory. Viscous wake effects were added. The displacement flow is clearly more important than the viscous wake flow. Both the numerical simulations and the model tests with rigid fish in current show that the fish influence on the mooring loads of the fish cage is less than 3% of the mooring load without fish. However, the measured mooring loads with live fish in current are between 10% and 28% larger than without fish. The reason is contact between the fish and the net cage. Accounting for the latter fact in the numerical model by changing the local solidity ratio of the net in the contact area gave reasonable numerical predictions. The experiments in waves and combined waves and current also showed a non-negligible influence of the fish on the mooring loads. The waves influenced the behaviour of the fish and some of the fish went to the net bottom possibly due to that they were uncomfortable in the wave zone.