τ Boo is an intriguing planet-host star that is believed to undergo magnetic cycles similar to the Sun, but with a duration that is about one order of magnitude smaller than that of the solar cycle. With the use of observationally derived surface magnetic field maps, we simulate the magnetic stellar wind of τ Boo by means of three-dimensional magnetohydrodynamics numerical simulations. As the properties of the stellar wind depend on the particular characteristics of the stellar magnetic field, we show that the wind varies during the observed epochs of the cycle. Although the mass-loss rates we find (∼2.7 × 10−12 M⊙ yr−1) vary less than 3 per cent during the observed epochs of the cycle, our derived angular-momentum-loss rates vary from 1.1 to 2.2 × 1032 erg. The spin-down times associated with magnetic braking range between 39 and 78 Gyr. We also compute the emission measure from the (quiescent) closed corona and show that it remains approximately constant through these epochs at a value of ∼1050.6 cm−3. This suggests that a magnetic cycle of τ Boo may not be detected by X-ray observations. We further investigate the interaction between the stellar wind and the planet by estimating radio emission from the hot Jupiter that orbits at 0.0462 au from τ Boo. By adopting reasonable hypotheses, we show that, for a planet with a magnetic field similar to Jupiter (∼14 G at the pole), the radio flux is estimated to be about 0.5–1 mJy, occurring at a frequency of 34 MHz. If the planet is less magnetized (field strengths roughly smaller than 4 G), detection of radio emission from the ground is unfeasible due to the Earth’s ionospheric cut-off. According to our estimates, if the planet is more magnetized than that and provided the emission beam crosses the observer line-of-sight, detection of radio emission from τ Boo b is only possible by ground-based instruments with a noise level of ≲1 mJy, operating at low frequencies.