In order to understand litter water-holding characteristics in secondary forests under different succession stages, three typically subtropical secondary forests were selected in Dashanchong Forest Park in Hunan Province. New litters were collected in every two months and measured its water holding capacity, water holding rate and water absorption rate by water soaking method. Then, the water holding capacity of litters in different forest types were compared. The relationships were analyzed between water holding capacity and carbon and nitrogen contents in litters. The results showed that:(1) the amounts and composition of litters in the three secondary forests presented unique changes. The amounts of litters in coniferous forest and evergreen broad-leaved forest were the largest from May to September in summer, while amounts of litters in deciduous broad-leaved forest were the largest in spring and autumn. (2) The saturated water holding capacity, half saturation time, and water affinity coefficients of litters in three secondary forests exhibited significantly seasonal variation characteristics. The saturated water holding capacity in coniferous forest litters reached a maximum of (59.68±2.91) g/m² from May to July, and the saturated water holding capacity in evergreen broad-leaved forest reached a maximum in September; the saturated water holding capacity of deciduous broad-leaved forest reached (190.60±8.81) g/m² in November. The half saturation time of litters in the three secondary forests was the lowest in November, and the half saturation time in deciduous broad-leaved forest litters was lower than that of other two forests types, with an annual average of (0.62±0.12) h; the litters' water affinity coefficient was 142.72±26.12 in deciduous broad-leaved forest in the whole year. (3) The saturated water holding rate of litters in the deciduous broad-leaved forest was significantly higher than that in the other types throughout the year (P<0.01). The saturated water holding rate of litters in coniferous forest and deciduous broad-leaved forest reached the maximum in November. (4) The water absorption rate coefficient A of litters in deciduous broad-leaved forest was significantly lower than that in the other types (P<0.01), while the water absorption rate coefficient B in coniferous forest litters was significantly higher than that in the other types (P<0.01). (5) The saturated water holding capacity of litters showed significantly positive correlations with water affinity coefficients, saturated water holding rate, the total amount of carbon and nitrogen in litters. The saturated water holding rate showed significantly negative correlation with half saturation time and water absorption rate coefficient A and B, significantly negative correlations with carbon content and C/N ratio in litters, and significantly positive correlations with nitrogen content in litters (P<0.01). In summary, the water holding capacity of litters in the three secondary forests were significantly different. The water holding capacity was significantly correlated with carbon and nitrogen content in litters. This study enriches the research theory of litter water-holding characteristics and provides a theoretical basis for in-depth discussion of the ecological and environmental effects of forest litters.