The acoustic nonlinearity parameter B/A describes the nonlinear features of a medium and may become a novel parameter for ultrasonic tissue characterization. This paper presents a theoretical analysis for acoustic nonlinear parameter tomography via a parametric array. As two primary waves of different frequencies are radiated simultaneously from a circular piston source, a secondary wave at the difference frequency is generated due to the nonlinear interaction of the primary waves. The axial and radial distributions of sound pressure amplitude for the generated difference frequency wave in the near field are calculated by a superposition of Gaussian beams. The calculated results indicated that the difference frequency component of the parametric array grows linearly with distance from the piston source. It therefore provides a better source to do the acoustic nonlinearity parameter tomography because the fundamental and second harmonic signals both have a near field that goes through many oscillations due to diffraction. By using a finite-amplitude insert substitution method and a filtered convolution algorithm, a computer simulation for B/A tomography from the calculated sound pressure of the difference frequency wave is studied. For biological tissues, the sound attenuation is considered and compensated in the image reconstruction. Nonlinear parameter computed tomography (CT) images for several biological sample models are obtained with quite good quality in this study.