We consider a 2D model of an autophoretic particle in which the particle has a circular shape and emits/absorbs a solute that diffuses and is advected by the suspending fluid. Beyond a certain emission/absorption rate (characterized by a dimensionless Péclet number, Pe) the particle is known to undergo a bifurcation from a non motile to a motile state, with different trajectories, going from a straight to circular and to a chaotic motion by progressively increasing Pe. From the full model involving solute diffusion and advection, we derive a reduced closed model which involves only two time-dependent amplitudes C1(t) and C2(t) corresponding to the first two Fourier modes of the solute concentration field. This model consists of two coupled nonlinear ordinary differential equations for C1 and C2 and presents several great advantages:(i) the straight and circular motions can be handled fully analytically, (ii) complex motions such as chaos can be analyzed numerically very efficiently in comparison to the numerically expensive full model involving partial differential equations, (iii) the reduced model has a universal form dictated only by symmetries, (iv) the model can be extended to higher Fourier modes. The derivation method is exemplified for a 2D model, for simplicity, but can easily be extended to 3D, not only for the presently selected phoretic model, but also for any model in which chemical activity triggers locomotion. A typical example can be found, for example, in the field of cell motility involving acto-myosin kinetics. This strategy offers an interesting way to cope with swimmers on the basis of ordinary differential equations, allowing for analytical tractability and efficient numerical treatment.