This paper addresses a peiformance analysis ofAffine Moment Invariants for 3D object recognition. Affine Moment Invariants are commonly used as shape feature for 2D object or pattern recognition. However, (his study proves that with some adaptation to multiple views technique, Affine Moment Invariants are sufficient to model 3D objects. In addition, the simplicity of moments calculation reduces the processing time for feature ex(raction, hence increases the system efficiency. In the recognition stage, this study used a neuro-fuzzy classifier called Multiple Adaptive Network based Fuzzy Inference System (MANFlS) for matching and classification. The proposed method was tested using two groups ofobject; polyhedral andfree-jorm objects. The experimental results show that Affine Moment Invariants combined with MANFIS nern:ork attain the best performance in both recognitions, polyhedral and free-form objects.