Aluminophosphate zeolites, including pure aluminophosphate (AlPO) and heteroatom-stabilized AlPO zeolites, have important applications in adsorption, separation and heterogeneous catalysis. Thus far, millions of hypothetical zeolite structures have been predicted, providing a large number of candidates to be synthetically targeted. However, their realization in experiment still requires a priori knowledge on whether heteroatoms are necessary in the synthetic preparation in order to stabilize a specific zeolite topology. To this end, many computational efforts have been made to compare the differences in framework energies and distortions before and after heteroatom incorporation. However, such an approach is not generally applicable for high-throughput computations, because of the combinatorial explosion of potential heteroatom incorporation sites in a hypothetical zeolite framework. Here, we establish a quantitative model to estimate the probability of a hypothetical framework being realizable as a pure AlPO or a heteroatom-stabilized AlPO zeolite. This model is based on Mahalanobis distances between hypothetical structures and their neighboring reference structures in distortion-energy plots. Our approach only requires onestep geometry optimization on zeolite frameworks as pure AlPO polymorphs without building many heteroatom-containing models, ensuring its applications in high-throughput structure evaluation and screening. We employed this approach on 84,292 hypothetical ABC-6 zeolite structures, and discovered that 17,050 of them could be realizable as pure AlPOs and 12,039 only realizable via heteroatom incorporation. Our results will provide important guidance toward the synthesis of new aluminophosphate zeolites.