Optimizing the oxygen content of silicon oxides used as anode materials for high-performance lithium-ion batteries is critical since it has diverse effects on lithium storage properties. However, the atomic-scale understanding of the effect of the oxygen content on structural evolution of the materials, particularly during delithiation is still limited. With this aim, we employ an iterative lithium extraction-relaxation algorithm based on reactive molecular dynamics simulations to investigate the dynamic processes of various silicon oxides. Our study highlights the effect of the oxygen content on various lithium storage properties, including the energetics of the interaction, lithium diffusion behaviors, lithium entrapment, structural reversibility during cycling, and mechanical properties. In particular, the simulations show that although increasing the oxygen content can significantly reduce the percentage of volume expansion during lithiation, which is a major cause of poor cycle retention of silicon- and silicon oxide-based anodes, this leads to an increased amount of the trapped lithium and an irreversible structural change during the reverse process. As contradictory conditions are required to achieve high specific capacity, long cycle life, and high coulombic efficiency, this atomistic study could provide a fundamental understanding beneficial for optimizing the oxygen content toward the development of silicon oxide-based anodes.