Baroreflex is commonly typified from heart period (HP) and systolic arterial pressure (SAP) spontaneous variations in the frequency domain mainly by estimating its sensitivity. However, an informative parameter linked to the rapidity of the HP response to SAP changes, such as the baroreflex bandwidth, remains unquantified. We propose a model-based parametric approach for estimating the baroreflex bandwidth from the impulse response function (IRF) of the HP-SAP transfer function (TF). The approach accounts explicitly for the action of mechanisms modifying HP regardless of SAP changes. The method was tested during graded baroreceptor unloading induced by head-up tilt (HUT) at 15°, 30°, 45°, 60°, and 75° (T15, T30, T45, T60, and T75) in 17 healthy individuals (age: 21–36 yr; 9 females and 8 males) and during baroreceptor loading obtained via head-down tilt (HDT) at −25° in 13 healthy men (age: 41–71 yr). The bandwidth was estimated as the decay constant of the monoexponential IRF fitting. The method was robust because the monoexponential fitting described adequately the HP dynamics following an impulse of SAP. We observed that 1) baroreflex bandwidth is reduced during graded HUT and this narrowing is accompanied by the reduction of the bandwidth of mechanisms that modify HP regardless of SAP changes and 2) baroreflex bandwidth is not affected by HDT but that of SAP-unrelated mechanisms becomes wider. This study provides a method for estimating a baroreflex feature that provides different information compared with the more usual baroreflex sensitivity while accounting explicitly for the action of mechanisms changing HP irrespective of SAP.