This study describes the reduction of the computer time requirements for an aerothermoelastic computational framework. The reduced order model (ROM) of the fluid solver is developed using kriging and proper orthogonal decomposition. The ROM accelerates the fluid solution by five orders of magnitude. In addition, a smart ordering scheme is proposed to accelerate the sample generation for the ROM. Using the extended computational framework, panel flutter problems are investigated with emphasis on flow orientation angle and boundary layer thickness. These parameters are found to have significant influence on the critical flutter dynamic pressure and amplitude of limit cycle oscillation.