Every smooth closed curve can be represented by a suitable Fourier sum as a function of an arbitrary parameter τ . We show that the ensemble of curves generated by randomly chosen Fourier coefficients with amplitudes inversely proportional to spatial frequency (with a smooth exponential cutoff) can be accurately mapped on the physical ensemble of inextensible worm-like polymer loops. The τ → s mapping of the curve parameter τ on the arc length s of the inextensible polymer is achieved at the expense of coupling all Fourier harmonics in a non-trivial fashion. We characterize the obtained ensemble of conformations by looking at tangent-tangent and position-position correlations. Measures of correlation on the scale of the entire loop yield a larger persistence length than that calculated from the tangent-tangent correlation function at small length scales. The topological properties of the ensemble, randomly generated worm-like loops, are shown to be similar to those of other polymer models. , use the analogy between conformation of a chain molecule and a Brownian random walk (BRW). In mathematically idealized form, such a random walk is thought of as a Wiener trajectory r(τ ) generated by the measure P {r(τ )} ∝ exp −const (∂r/∂τ ) 2 dτ , where τ is a parameter running along the trajectory. Despite its successes, the BRW model cannot reproduce the finite extensibility of polymer chains stretched by a strong force [2] and fails miserably in the study of polymers with knots. Indeed, simulations of discrete polymer models, or random polygons with N steps, show [3][4][5] that the probability of trivial knot configuration decays exponentially with N as P 0 (N) ∼ exp(−N/N * ), where N * defines the crossover from an unknotted to a knotted regime. It was argued that Wiener trajectory polymer models cannot be used to calculate