Producing oxygen with purity higher than 95.0% from atmospheric air (78.0% N-2, 21.0% O-2, and 1.0% Ar) is challenging because of the similar physical properties of oxygen and argon. Silver-exchanged titanosilicates have shown the potential to separate these gases based on their thermodynamic affinities. In this study, various vacuum swing adsorption (VSA) cycle configurations including the simple Skarstrom cycle and more complicated 6-step VSA cycles were simulated using mathematical models to maximize O-2 purity and recovery. The simulations were verified by conducting simple 3-step and Skarstrom VSA cycle experiments. A mixture of 95.0%/5.0% O-2/Ar feed was considered in the simulations, and a rigorous multiobjective optimization was conducted to maximize O-2 purity and recovery. The simulations predicted 27.3% recovery for a product with 99.5% purity for a 6-step cycle with pressure equalization and light product pressurization steps. The recovery for the same level of purity was improved significantly to 91.7% by implementing a heavy product pressurization step. The effect of bed length on O-2 purity and recovery and the comparison of VSA with pressure swing adsorption and pressure vacuum swing adsorption for high-purity O-2 production were also investigated. Rigorous multiobjective optimizations were conducted to maximize oxygen productivity and minimize energy consumption of the VSA cycles, while meeting different purity constraints, and significant improvement in the performance indicators was obtained through process optimization.