We explore the wind braking of magnetars considering recent observations challenging the traditional magnetar model. There is evidence for strong multipole magnetic fields in active magnetars, but the dipole field inferred from spin-down measurements may be strongly biased by particle wind. Recent observations challenging the traditional model of magnetars may be explained naturally by the wind braking scenario: (1) the supernova energies of magnetars are of normal value; (2) the non-detection in Fermi observations of magnetars; (3) the problem posed by low magnetic field soft gamma-ray repeaters; (4) the relation between magnetars and high magnetic field pulsars; and (5) a decreasing period derivative during magnetar outbursts. Transient magnetars with L-x<-(E) over dot(rot) may still be magnetic dipole braking. This may explain why low luminosity magnetars are more likely to have radio emissions. A strong reduction of the dipole magnetic field is possible only when the particle wind is very collimated at the star surface. A small reduction of the dipole magnetic field may result from detailed considerations of magnetar wind luminosity. In the wind braking scenario, magnetars are neutron stars with a strong multipole field. For some sources, a strong dipole field may no longer be needed. A magnetism-powered pulsar wind nebula will be one of the consequences of wind braking. For a magnetism-powered pulsar wind nebula, we should see a correlation between the nebula luminosity and the magnetar luminosity. Under the wind braking scenario, a braking index smaller than three is expected. Future braking index measurement of a magnetar may tell us whether magnetars are wind braking or magnetic dipole braking.