We report on very-long-baseline-interferometry (VLBI) monitoring observations of the low-luminosity active galactic nucleus (LLAGN) in the galaxy M 81 at the frequencies of 1.7, 2.3, 5.0, and 8.4 GHz. The observations reported here are phase-referenced to the supernova SN 1993J (located in the same galaxy) and cover from late 1993 to late 2005. The large amount of available observations allows us to study the stability of the AGN position in the frame of its host galaxy at different frequencies and chromatic effects in the jet morphology, together with their time evolution. The source consists at all frequencies of a slightly resolved core and a small jet extension towards the northeast direction (position angle of ∼65 degrees) in agreement with previous publications. We find that the position of the intensity peak in the images at 8.4 GHz is very stable in the galactic frame of M 81 (proper motion upper limit about 10 μas per year). We confirm previous reports that the peaks at all frequencies are systematically shifted among them, possibly due to opacity effects in the jet as predicted by the standard relativistic jet model. We use this model, under plausible assumptions, to estimate the magnetic field in the jet close to the jet base and the mass of the central black hole. We obtain a black-hole mass of ∼2 × 10 7 M , comparable to estimates previously reported using different approaches, but the magnetic fields obtained are 10 3 −10 4 times lower than previous estimates. We find that the positions of the cores at 1.7, 2.3, and 5.0 GHz are less stable than that at 8.4 GHz and evolve systematically, shifting southward at a rate of several tens of μas per year. The evolution in the jet orientation seems to be related to changes in the inclination of the cores at all frequencies. These results can be interpreted as due to a precessing jet. The evolving jet orientation also seems to be related to a flare in the peak flux densities at 5.0 and 8.4 GHz, which lasts ∼4 years (from mid 1997 to mid 2001). An increase in the accretion rate of the black hole, and its correlation with the jet luminosity via the disk-jet connection model, seems insufficient to explain this long flare and the simultaneous evolution in the jet orientation. A continued monitoring of the flux density and the jet structure evolution in this LLAGN will be necessary to further confirm our jet precession model.