The aim of the study was epidemiological analysis of survival from all types of leukemia occurring in Cracow in the years 1980-1990. The study was focused on survival times in patients according to a) cytologico-clinical type of leukemia, b) timeframe in which treatment was initiated (between 1980-1985 and 1986-1090). All patients diagnosed of leukemia between the years 1980-1990, living in Cracow and whose cytologico-clinical picture was determined had their survival times and censored survival times established. Survival until 1997 was taken into account. For each cytologico-clinical type of leukemia survival function according to Kaplan-Meier was calculated. The Cox model was implemented to analyze the risk of death depending on the period in which the disease appeared--two time frames were established 1980-1985 and 1986-1990. Other parameters considered were; age, sex and area in which the patient lived (suburb). Practically in all types of leukemia a higher probability of survival was found in patients in whom leukemia was diagnosed (and consequently treated) in the second period i.e., 1986-1990. The highest achievement was observed in acute lymphoblastic leukemia in children, in which the relative 5-year survival probability rose from 35% in the years 1980-1985 to 78% in the years 1986-1990, thus achieving the level of well developed countries. A similar picture was seen in chronic lymphocytic leukemia where the relative 5 year survival probability rose from 57% to 77%, and in chronic granulocytic leukemia where the 5 year survival probabilities were accordingly 23% and 39%. All cited values for the second period of analysis are at the levels noted in the United States and in Europe. The positive changes in the survival times observed in patients with leukemia seen in the second half of the 80-ies (in comparison to the period 1980-1985) has been interpreted as the result of advancements in therapy of the disease in Cracow.