Numerical simulations and an optimization method are used to study the design of a planar T-micromixer with curved-shaped baffles in the mixing channel. The mixing efficiency and the pressure loss in the mixing channel have been evaluated for Reynolds number (Re) in the mixing channel in the range 1 to 250. A Mixing index (Mi) has been defined to quantify the mixing efficiency. Three geometric dimensions: radius of baffle, baffles pitch and height of the channel, are taken as design parameters, whereas the mixing index at the outlet section and the pressure loss in the mixing channel are the performance parameters used to optimize the micromixer geometry. To investigate the effect of design and operation parameters on the device performance, a systematic design and optimization methodology is applied, which combines Computational Fluid Dynamics (CFD) with an optimization strategy that integrates Design of Experiments (DOE), Surrogate modeling (SM) and Multi-Objective Genetic Algorithm (MOGA) techniques. The Pareto front of designs with the optimum trade-offs of mixing index and pressure loss is obtained for different values of Re. The micromixer can enhance mixing using the mechanisms of diffusion (lower Re) and convection (higher Re) to achieve values over 90%, in particular for Re in the order of 100 that has been found the cost-effective level for volume flow. This study applies a systematic procedure for evaluation and optimization of a planar T-mixer with baffles in the channel that promote transversal 3-D flow as well as recirculation secondary flows that enhance mixing.