Concentrate of nanofiltration and reverse osmosis installations is an increasing problem, especially for inland membrane installations. Introduction of membrane filtration in The Netherlands is severely hindered by the concentrate problem. Two approaches are viable for solving or reducing the concentrate problem: (1) low recovery NF/RO without anti-sealant dosing, (2) Zero liquid discharge. This research focuses on the second option: Zero liquid discharge NF/RO. First and main problem to be solved with zero liquid discharge is to increase the recovery of the membrane installation to its limits, without increasing the costs of water produced. A high recovery (>99%) is necessary to reduce energy consumption and costs for evaporation of the remaining waste stream (<1%). The only possibility to achieve a very high recovery in NF/RO, is removing the scaling components from the feed water. A very important advantage of removing the scaling components is that the nanofiltration or RO can be operated at high fluxes. In this paper the results of two pilot experiments are reported. One treatment concept was developed for surface water treatment and one for groundwater treatment. The surface water treatment concept consisted of fluidized ion exchange to remove positive multivalent cations, and then followed by ultrafiltration, nanofiltration and granular activated carbon filtration. With this setup a recovery of 97% was achieved. To achieve an even higher recovery it is also important to remove silica from the feed water because silica can limit the recovery. Silica can be removed at high pH during co-precipitation with magnesium hydroxide. The groundwater concept consisted of: precipitation at high pH, then followed by sedimentation, weak acid cation exchange and nanofiltration. With this setup a recovery of 99% was achieved.