1. Genetic parameters in the base population of a closed experimental strain of Muscovy ducks, selected for body weight at 10 weeks of age, were estimated from data in 8 successive generations, for the following traits: age at first egg (AGE1EGG), total number of eggs laid at 40 and 52 weeks of age (NEGG40 and NEGG52), number of eggs laid during 15 and 22 weeks in the first laying cycle (NEGG15W and NEGG22W), and their Box-Cox transformed data. 2. The method of multi-trait restricted maximum likelihood with an animal model was used to estimate genetic parameters. Only the results obtained with non-transformed data are shown. 3. Heritability estimates for laying performance showed moderate values, increasing little with age: 0.20±0.03 (AGE1EGG), 0.23±0.03 (NEGG40), 0.27±0.03 (NEGG52), 0.20±0.03 (NEGG15W), and 0.22±0.03 (NEGG22W). 4. Genetic correlations between laying traits were high. Genetic correlation between AGE1EGG and egg number was negative, it was positive between total numbers of eggs at 40 and 52 weeks and egg numbers in the first laying cycle. 5. Body weight at 10 weeks of age exhibited positive genetic correlations (0.46±0.06) with age at first egg and negative with egg production traits (- 0.28±0.06 to - 0.41±0.06). 6. The cumulated predicted genetic gains, after 7 generations of selection, expressed per genetic standard deviation unit ([sigma]g) were 0.06[sigma]g, 0.07[sigma]g, 0.17[sigma]g, 0.23[sigma]g, and 0.25[sigma]g for AGE1EGG, NEGG40, NEGG52, NEGG15W, and NEGG22W, respectively. 7. Selecting Muscovy ducks to improve laying in Taiwanese climatic conditions would be possible using the number of eggs laid up to 52 weeks of age as the selection criterion. Because unintended selection effects for laying traits were present, the selection experiment for body weight at 10 weeks of age was not antagonistic with laying traits.