Conventional glasshouse climate controllers are based upon continuous-time PI controllers manually tuned to achieve adequate, although rather poor, tracking of set point changes. In this paper we consider the alternative, model based, Proportional-Integral-Plus (PIP) control system design (Young et al, 1987) which, although only slightly more complex than a PI controller, achieves much tighter control of the climate variables, allowing optimal setpoints (Chalabi, 1992) to be realised. A linear control model is identified and estimated from experimental data collected in a Venlo glasshouse at Silsoe Research Institute (SRI). A Non Minimum State Space (NMSS) representation of this control model is then used to design a robust PIP controller which was implemented during the 1993/94 winter growing season with a tomato crop. Control results were excellent with very tight control to the desired setpoints in all three variables. Air temperature was controlled to within 0.5°C of the setpoint for 85% of the validation period, and was shown to be very robust to model uncertainty and extreme weather conditions. Relative humidity was controlled to within 2% RH for 90% of the validation period, and CO2 was controlled to within 15 ppm for 80% of the validation period.