Cao, J. and McMechan, G.A., 2011. Free-surface multiple prediction and subtraction from slowness relations in 2D and 3D synthetic data. Journal of Seismic Exploration, 20: 235-255. A target-oriented, data-adaptive, algorithm is developed for the prediction and subtraction of free-surface multiples from seismic data, without knowledge of the subsurface velocities. It uses only slowness relations between the primary reflections on split-spread common-source and common-receiver gathers. It is based on matching slownesses at all source and receiver locations and combining offsets and times of primary reflections to kinematically predict multiples. Our use of only the slownesses p in the multiple prediction eliminates many of the assumptions and complexities that are involved in previous algorithms. This method is extended to 3D and to predict all higher-order multiples. The inputs are the traveltimes, of the primary reflections that produce the multiples, picked from common-source gathers. The subtraction involves flattening the multiple events on their predicted traveltime trajectories, estimating and subtracting a local spatial average trace from the center trace in a moving trace window, and then shifting each trace back to its original time. The effectiveness of this algorithm is illustrated using 2D and 3D synthetic examples. Multiple reduction is clearly visible in common-source and common-offset sections, before and after prestack migration.