Noises and stochastic effects exist in most biological systems due to many intrinsic and extrinsic factors. Here I will discuss strategies and principles for noise attenuation and robustness to genetic or/and environmental perturbations in cell signaling and embryonic patterning. In one case, I will introduce a critical quantity that dictates capability of attenuating temporal noises in feedback systems. In another case, I will show that noises in signal transduction surprisingly enable reduction of stochastic effects and uncertainty in developmental patterning. For both cases, several new experimental data (from yeast cells and zebrafish embryos) that support modeling and computational predictions will be presented. In addition, mathematical and computational challenges associate with such study will be discussed.