This paper examines the ability of the dynamic optimization model to predict changes between in-phase and anti-phase postural modes of coordination and to evaluate influence of two particular environmental and intentional constraints on postural strategy. The task studied was based on an experimental paradigm that consisted in tracking a target motion with the head. An original optimal procedure was developed for cyclic problems to calculate hip and ankle angular trajectories during postural sway with a minimum torque change criterion. Optimization results give a good description of the sudden bifurcation phase between in-phase and anti-phase postural coordination modes in visual target tracking. Transition frequency and predicted effects of environmental and intentional constraints are also in line with experimental observations described in existing literature. In particular, these investigations pointed out that postural planning process can be related to the minimization of a dynamic cost criterion with an equilibrium constraint. In conclusion, the optimization technique is well suited for the prediction of postural modes of coordination and seems to offer many opportunities for better comprehension of neuromuscular movement control.