We assessed the association between usual coffee consumption and all-cause, cardiovascular (CV), and cancer mortality in an adult population in Spain, taking into account both the amount and type of coffee consumed. We used baseline data on coffee consumption and other personal variables, and the number of deaths during an 18-year follow-up period, for 1567 participants aged 20 years and older from the Valencia Nutrition Study in Spain. Total, caffeinated, and decaffeinated coffee consumption was assessed using a validated food frequency questionnaire. Cox regression models were used to estimate adjusted hazard ratios (HRs) and 95% confidence intervals (CIs). During the 18-year follow-up period, 317 died; 115 due to CV disease and 82 due to cancer. Compared with no-consumption, the consumption of <= 1 cup per day and >1 cup per day of coffee was associated with a lower risk of all-cause mortality, HR = 0.73 (95% CI: 0.56-0.97) and HR 0.56 (95% CI: 0.41-0.77), respectively. A lower cancer mortality was observed among drinkers of more than 1 cup per day compared with nondrinkers, HR 0.41 (95% CI 0.20-0.86). Regarding the type of coffee, only the overall consumption of caffeinated coffee was associated with lower all-cause mortality at 12 and 18 years of follow-up, HR = 0.66 (95% CI:0.46-0.94) and HR = 0.59 (95% CI: 0.44-0.79), respectively. In conclusion, this study suggests that the moderate consumption of coffee, particularly caffeinated coffee (range 1-6.5 cups per day), is associated with a lower all-cause and cancer mortality after a long follow-up period. No significant association was found between coffee consumption and CVD mortality.