We present a consistent, physically based theory for the wake meandering phenomenon, which we consider of crucial importance for the overall description of wind turbine loadings in wind farms. In its present version the model is confined to single wake situations. The model philosophy does, however, have the potential to include also mutual wake interaction phenomenons. The basic conjecture behind the dynamic wake meandering model is that wake transportation in the atmospheric boundary layer is driven by the large scale lateral- and vertical turbulence components. Based on this conjecture a stochastic model of the downstream wake meandering is formulated. In addition to the kinematic formulation of the dynamics of the “meandering frame of reference”, models characterizing the mean wake deficit as well as the added wake turbulence, described in the meandering frame of reference, are an integrated part the model complex. For design applications, the computational efficiency of wake deficit prediction is a key issue. Two computationally low cost models are developed for this purpose. The character of the added wake turbulence, generated by the up-stream turbine in the form of shed and trailed vorticity, has been approached by analytical as well as by numerical studies. The dynamic wake meandering philosophy has been verified by comparing model predictions with extensive full-scale measurements. These comparisons have demonstrated good agreement, both qualitatively and quantitatively, concerning both flow characteristics and turbine load characteristics. Contrary to previous attempts to model wake loading, the dynamic wake meandering approach opens for a unifying description in the sense that turbine power– and load aspects can be treated simultaneously. This capability is a direct and attractive consequence of the model being based on the underlying physical process, and it potentially opens for optimization of wind farm topology, of wind farm operation as well as of control strategies for the individual turbine. Implementation of the methodology with aeroelastic codes is straight forward and performed simply by replacing traditional turbulence input files with wind field files containing the combined effect of atmospheric turbulence and wake meandering.