The problem of parametric modeling for the two-dimensional (2-D) harmonic signals with missing harmonics is addressed. The modeling process consists of two parts. First, we devise the joint spatial-temporal linearly constrained minimum variance beamformer and perform the joint estimation of the spatial and temporal fundamental frequencies for each pitch of the signal based on the maximum harmonic model. Second, we differentiate the competing signal models by the number of harmonic components and derive the maximum a posteriori criterion for the 2-D harmonic signals. As a result, the harmonic components of each pitch are detected according to their spectral powers. The simulation and experimental results are provided to show the superiority of the proposed signal modeling methodology over other existing schemes.