In-duct sorbent injection (DSI) is a well-known, low-cost desulphurisation technology handicapped by its moderate SO2 removal capacity. Fortunately, there are some technical options for increasing the desulphurisation efficiency without eliminating its inherent advantages. In this ex perimental study, several improvement design options like the recirculation of reactivated sorbent, the pre-collection of the fly ash and the use of seawater for humidification have been analysed using an extensive parametric testings programme. The effect of the main operating variables directly related to the desulphurisation efficiency has been also tested following a fractional factorial design. These variables were the Ca/S ratio, the approach to the adiabatic saturation temperature and the recirculation ratio of the partially converted sorbent. Other important questions like the use of a high-BET-area lime and the impact of the DSI process on an ESP have been also included in this experimental assessment. More than 50 experimental tests were carried out in a 3-MWe equivalent pilot plant to assess the different improvement options for in-duct sorbent injection. The results of this study allow us to extract practical conclusions about the devices, equipment and operating conditions as a function of the target SO2 efficiency, and even enable us to provide an economic assessment. Using the proposed improvement options to process a flue gas with 40-1000 ppm of SO2 concentration, a 90% sulphur removal with a lime utilisation of 45% was achieved.