The reflection of elastic waves from interfaces is a robust phenomenon extensively used by animals (bats and dolphins), nondestructive techniques, and medical diagnosis. This study addresses the design and implementation of P-wave reflection imaging to evaluate the internal variability in small-scale submerged, soil models. The performance of this technology depends on fundamental aspects of P-wave propagation in soils, the selection of optimal P-wave transducers, and the development of an adequate test methodology. Design issues include transducer directivity, noise, axial resolution, near field effects, and proper thickness of the coupling water layer. The operating frequency is determined by transducer selection and affects the axial and lateral resolution, skin depth, near field, and divergence; high damping transducers permit higher axial resolution. In addition, data gathering must take into consideration temporal and spatial aliasing. Results show that P-wave reflection is a valuable tool to detect subsurface anomalies and layers, to assess phenomena such as slurry sedimentation, and to monitor the evolution of subsurface structures such as soil layers during liquefaction. Gradual changes in impedance, such as in slurry sedimentation, may prevent reflections.