This talk deals with two main challenging issues in humanoid robotics, namely pattern generation and control design. The presentation includes three main parts; the first one is an introduction to humanoid robotics, where their basic principles are introduced. The second part deals with the problem of pattern generation in humanoid robotics. It starts with the problem formulation of pattern generation, followed by some proposed solutions (such as the three-mass linear inverted pendulum model, B-spline based and human data based generators) and their validations. The third part of the presentation concerns control design of humanoid walking robots. First the problem formulation of control design is introduced, and then the proposed solutions are presented. The first solution deals with control of two-leg (with only lower limbs) biped walking robots. The proposed control architecture includes a ZMP-based pattern generator, an inverse dynamics controller, and a stabilizer to ensure dynamic walking stability. The proposed solution is validated through numerical simulations on the walking robot SHERPA. The second solution, concerns whole-body control, and is based on human walking motion capture. This technique is a very interesting concept in design of human-like walking; where mainly articular positions are used to create the human motion database. However, this method needs a huge amount of data. Therefore, a reduction of the data set is necessary to keep only the most valuable information; furthermore this can help for a better understanding of human walking and a smarter implementation of human-like walking on humanoid robots. In our study, we take into account the above mentioned fact concerning data reduction. Consequently, only the feet and center of mass positions are needed to describe the important features of human movement. Then the redundancy of the humanoid robot is considered to track these two objectives using the task formalism. The proposed solution is validated by numerical simulations as well as real-time experiments on the humanoid robot HOAP3.