Diffusion Probabilistic Models (DPMs) have shown significant performance in high-resolution image synthesis, but their sampling efficiency remains a challenge due to the large number of sampling steps required. Recent advancements in high-order numerical ODE solvers have improved this efficiency, but most methods still use uniform time steps, which is suboptimal with fewer steps. To address this, the authors propose a general framework for designing an optimization problem to find more appropriate time steps for specific numerical ODE solvers in DPMs. This optimization problem aims to minimize the distance between the ground-truth solution and an approximate solution. The problem can be efficiently solved using the constrained trust region method, taking less than 15 seconds. Extensive experiments on unconditional and conditional sampling using pixel- and latent-space DPMs demonstrate that the optimized time steps significantly improve image generation performance, as measured by FID scores, compared to uniform time steps.Diffusion Probabilistic Models (DPMs) have shown significant performance in high-resolution image synthesis, but their sampling efficiency remains a challenge due to the large number of sampling steps required. Recent advancements in high-order numerical ODE solvers have improved this efficiency, but most methods still use uniform time steps, which is suboptimal with fewer steps. To address this, the authors propose a general framework for designing an optimization problem to find more appropriate time steps for specific numerical ODE solvers in DPMs. This optimization problem aims to minimize the distance between the ground-truth solution and an approximate solution. The problem can be efficiently solved using the constrained trust region method, taking less than 15 seconds. Extensive experiments on unconditional and conditional sampling using pixel- and latent-space DPMs demonstrate that the optimized time steps significantly improve image generation performance, as measured by FID scores, compared to uniform time steps.