This paper is concerned with diffusive approximations of peculiar numerical schemes for several linear (or weakly nonlinear) kinetic models which are motivated by wide-range applications, including radiative transfer or neutron transport, run-and-tumble models of chemotaxis dynamics, and Vlasov-Fokker-Planck plasma modeling. The well-balanced method applied to such kinetic equations leads to time-marching schemes involving a "scattering S-matrix", itself derived from a normal modes decomposition of the stationary solution. One common feature these models share is the type of diffusive approximation: their macroscopic densities solve drift-diffusion systems, for which a distinguished numerical scheme is Il'in/Scharfetter-Gummel's "exponential fitting" discretization. We prove that the well-balanced schemes relax, within a parabolic rescaling, towards the Il'in exponential-fitting discretization by means of an appropriate decomposition of the S-matrix. This is the so-called asymptotic preserving (or uniformly accurate) property.Obviously, such a general statement contains several former ones, among which the two-stream Goldstein-Taylor model relaxing to the heat equation, [29], or the so-called "Cattaneo model of chemotaxis", [22]; some of these results were surveyed in [24, Part II].Yet, as a guideline for more involved calculations, we first explain in §3 how the limiting process works for a simple two-stream approximation of (1.1), the so-called "Greenberg-Alt" model of chemotaxis [32]. We mention that for this simple two-velocity model, a numerical scheme formulated in terms of (2.10) with 2 × 2 S-matrices is provided in both [24, page 158] and [30, Lemma 4.1] (for the purpose of hydrodynamic limits, though). Another type of closely related "diffusive limit" involving a 2 × 2 S-matrix was studied in [27].This elementary calculation carried out on the two-stream "Greenberg-Alt" model (3.18) reveals why it is rather natural to expect that a well-balanced algorithm (3.24) (based on stationary solutions) may relax, within a parabolic scaling, toward the exponential-fit scheme (3.21)-(3.22) for the corresponding asymptotic Keller-Segel model. However, as our Theorem 1.1 covers also continuous-velocity models discretized with general quadrature rules, we present in §2 our strategy of proof: in particular, the general scheme involving a scattering matrix is presented in (2.10) and the importance of the decomposition of the scattering matrix (2.11) is emphasized. In §4, such a strategy is applied to the simplest case of continuous equation, namely the "grey radiative transfer" model (4.29). For this system, it is shown in Theorem 4.6 that our numerical scheme relaxes to the finite-difference discretization of the heat equation (4.55). In §5, the case of the Othmer-Alt [44] model of chemotaxis dynamics (5.56)-(5.57) is handled in a similar manner (at the price of more intricate computations, though), yielding asymptotically the scheme (5.59), this is Theorem 5.5. At last, in §6, the case of a Vlasov-Fokker-Planck model ...