The Kolmogorov–Sinai (KS) entropy in turbulent diffusion of magnetic field lines is analyzed on the basis of a numerical simulation model and theoretical investigations. In the parameter range of strongly anisotropic magnetic turbulence the KS entropy is shown to deviate considerably from the earlier predicted scaling relations (1992 Rev. Mod. Phys.64 961). In particular, a slowing down logarithmic behavior versus the so-called Kubo number R ≫ 1 (R = (δB/B0) (ξ∥/ξ⊥), where δB/B0 is the ratio of the rms magnetic fluctuation field to the magnetic field strength, and ξ⊥ and ξ∥ are the correlation lengths in respective dimensions) is found instead of a power-law dependence. These discrepancies are explained from general principles of Hamiltonian dynamics. We discuss the implication of Hamiltonian properties in governing the paradigmatic 'percolation' transport, characterized by R → ∞, associating it with the concept of pseudochaos (random non-chaotic dynamics with zero Lyapunov exponents). Applications of this study pertain to both fusion and astrophysical plasma and by mathematical analogy to problems outside the plasma physics.