Fish larvae have the ability to change their vertical position in the water column and thusly cannot be treated as passive particles in coupled biological-physical individualbased models (IBMs). The vertical variability of light, turbulence, temperature, prey, predators, and horizontal currents in the ocean affects the survival of larval fish through effects on feeding, growth, advection, and predation mortality. A dynamic model of the vertical position of larval fish in response to individual state and environmental conditions is needed for use in three-dimensional IBMs. A 1-dimensional model was constructed of an idealized water column representative of spring conditions on the southern flank of Georges Bank. The water column was used to test six behavioral rules of individuals parameterized as larval haddock (Melanogrammus aeglefinus) under different conditions of prey and turbulence stratification. Our objectives were to determine how behaviors based on different state and environmental variables affect depth distribution and mortality, and which behaviors produce a vertical distribution most similar to observations. Individuals applying behaviors associated with feeding had distributions comparable to observations and the highest survival. The use of behaviors derived from a trade-off between gut fullness and visual predation led to distributions unlike observations and high starvation mortality of the largest larval size class. Results suggest that larvae should make their vertical behavior decisions based on the risk of starvation rather than predation. A realistic model of larval haddock vertical position could be developed using only behaviors related to its prey distribution and foraging success.