Eight trials were conducted in commercial potato fields infested with the white potato cyst nematode (wPCN, Globodera pallida) and one in a field infested with the yellow PCN (yPCN, Globodera rostochiensis). Our aims were to produce data to validate and refine a computer-based program (The Model) for the long-term management of PCN, to determine nematicide effectiveness and to assess rates of PCN population decline between potato crops. Prior to planting, each farmer applied an overall nematicide treatment to his field, except for ten untreated plots that were widely spaced to encompass a range of PCN population densities. Each untreated plot was paired with a similar plot in the adjacent treated area and all plots were intensively sampled for PCN population densities at planting (Pi) and again at harvest (Pf) when tuber yields were determined. Four trials were re-sampled 2–4 years later to determine PCN population decline rates. Regressions that form the basis of ‘The Model’ and described the relationship between Pi and tuber yield and PCN population density at harvest were fitted to the results from both the untreated and nematicide treated plots. These regressions also enabled us to estimate the yield potential at each site in the absence of PCN and showed that nematicide treatment generally did not increase yield potential and that both tuber yield and PCN multiplication decreased with increasing Pi. However, there were major differences between sites and cultivars. When untreated, the yield of cv. Maris Piper was hardly affected in a highly organic soil with Pi > 200 eggs g−1 whereas the yield of partially resistant cv. Santé was decreased from a potential of c. 60 t ha−1 to c. 20 t ha−1 in a light silt with Pi = 20 egg g−1 soil. Similarly, untreated wPCN multiplication rates at a low Pi ranged from 46-fold to >100-fold. Nematicide effectiveness was estimated from the regressions and, at several sites, yield was decreased despite nematicide treatment. Control of wPCN multiplication was even poorer. In only two of seven trials planted with susceptible cultivars was more than 50% control achieved – maximum populations in treated plots usually exceeded 250 eggs g−1. Partially resistant Santé decreased the multiplication rate of wPCN in the two trials where it was planted. An alternative analysis using Genstat indicated that The Model tended to underestimate the maximum multiplication rate and overestimate the maximum population density. When four sites were re-sampled 2–4 years after harvest the populations of wPCN had declined by between 15% and 33.5% per annum with a mean of 26% per annum. Modelling indicated that rotations longer than 8 years were required to control wPCN unless other effective control measures, such as growing a partially resistant cultivar, were used.