Protodyakonov’s pressure arch theory (PPAT) has been widely used in rock engineering design owing to its conciseness and clear engineering concept. However, the applicability of this theory to soil formations has not yet reached a consensus, mainly because of the lack of criterion for stable arching in soil and the fact that soil has a much lower strength than rocks. To fill this gap, the PPAT based on arch assumptions and arch stability requirements is revisited in this paper. A modified arch equation (M-PPAT) is established considering the lateral stress and arch foot resistance. Expressions for the arch strength and arch foot stability criteria are given. A trapdoor test and PFC2D simulation is conducted to verify the M-PPAT. The difference between the arch height calculated by M-PPAT and the trapdoor test result is only 3.3%, which is much better than the height given by PPAT. The particle flow simulation results for the trapdoor test by PFC2D show that the particle material can form a stable naturally balanced arch (NBA), and the shape of the arch is very close to the test results. Taking typical rock and soil formations as examples, the inferences obtained by M-PPAT are discussed. The arch foot stability criterion was used to test the stability of the arches formed in the soil using the arch axis equation, which does not consider the arch foot stability, and it was found that these arches are unstable except for M-PPAT. The arch height of stable NBA is limited by the arch foot stability criterion, and the maximum buried depth for stable NBA was found to be limited by the arch strength criterion. Since M-PPAT does not consider cohesion, this theory should be used with caution in cohesive soil formations. The research results can be used to determine the applicable scope of the surrounding earth pressure calculation method that is based on the PPAT. Moreover, they have reference value for designing underground space structures.