A statistical model is developed to reconstruct the total ozone field over Europe since 1 January 1950. Daily total ozone data are required for surface UV reconstruction. The model is trained on satellite data (National Institute of Water and Atmosphere Research (NIWA) assimilated total column ozone database) over the period 1979–2004. The model explains ∼80% of variance of the satellite total ozone. Standard deviation of the relative differences, observed minus model daily data in percent of observed data, is ∼4% and ∼6% in May-September and October-next year April season, respectively. The total ozone values are reconstructed backward in time using the regression constants derived as a result of the training, time series of various atmospheric indices of the atmospheric circulation, and the meteorological variables. The proxies for the ozone variations are taken from the National Centers for Enivronmental Prediction (NCEP)/National Center for Atmosphere Research (NCAR) reanalysis database. Optimal combination of the proxies is selected using MARS technique. The quality of the database is assured by a comparison of the reconstructed total ozone with the ground-based data from several Dobson stations functioning in the early 1950s and 1960s. The model explains ∼70–80% variance of the ozone data collected before the satellite era. Bias and the long-term drift between the reconstructed and measured Dobson ozone are within a range of ±2%. Standard deviation of relative differences between reconstructed and station Dobson daily data is 5–7% and ∼10% in May-September and October-next year April season, respectively. The long-term tendency, comprising an anthropogenic trend and superposition of “natural” long-term fluctuations, appears almost trendless in the period 1950–1980 with a clear ozone decline after 1980.