A stack of records becomes one of the main steps in modern seismic data processing. In the stack procedure, the crucial operation is time correction. Conventional methods, e.g., normal moveout (NMO) and dip moveout (DMO) stacks require a sufficiently accurate macro-velocity model, whereas a multifocusing imaging method does not depend on a macro-velocity model. The multifocusing method proposed by Gelchinsky et al. belongs to a group of methods that can be characterized as macro-model-independent imaging methods. The multifocusing method represents a transformation of 2-D multicoverage reflection data into a simulated zero-offset stack profile. This transformation is based on a completely data-derived spatial stacking operator, and includes stacking large supergathers of seismic traces, each of which can span many CMP gathers. By extending the multifocusing moveout formula to explicitly account for non-zero elevations of the source and receiver, the multifocusing imaging method can yield appropriate results when seismic data are acquired over an irregular topography. In recent years, many applications of multifocusing imaging over an irregular topography have demonstrated its advantages in comparison with conventional CMP processing. This paper illustrates the corresponding formulas for a synthetic data example modeled by the wave equation finite difference method. The result of the synthetic example is very encouraging. By stacking large supergathers and applying multifocusing moveout correction, the reflectors are aligned very well and the S/N is greatly improved. We have also applied multifocusing imaging over an irregular topography to a real data example. The elevation of the data acquisition area varies considerably. Applying multifocusing imaging, a substantial improvement of the simulated section was achieved, compared with a conventional CMP stacked section.