Industrial glass blowing is an essential stage of manufacturing glass containers, i.e., bottles or jars. An initial glass preform is brought into a mold and subsequently blown into the mold shape. Over the past few decades, a wide range of numerical models for forward glass blow process simulation has been developed. A considerable challenge is the inverse problem: to determine an optimal preform from the desired container shape. A simulation model for blowing glass containers based on finite element methods has previously been developed (Giannopapa, 2008, “Development of a Computer Simulation Model for Blowing Glass Containers,” ASME J. Manuf. Sci. Eng., 130, p. 041003; Giannopapa and Groot, 2007, “A Computer Simulation Model for the Blow-Blow Forming Process of Glass Containers,” 2007 ASME Pressure Vessels and Piping Conference and 8th International Conference on CREEP and Fatigue at Elevated Temperature). This model uses level set methods to track the glass-air interfaces. The model described in a previous paper of the authors showed how to perform the forward computation of a final bottle from the given initial preform without using optimization. This paper introduces a method to optimize the shape of the preform combined with the existing simulation model. In particular, the new optimization method presented aims at minimizing the error in the level set representing the glass-air interfaces of the desired container. The number of parameters used for the optimization is restricted to a number of control points for describing the interfaces of the preform by parametric curves, from which the preform level set function can be reconstructed. Numerical applications used for the preform optimization method presented are the blowing of an axisymmetrical ellipsoidal container and an axisymmetrical jar.