This paper presents a method for capturing and efficiently simulating of a large number of plant leaves, realistically showing their weathering process in autumn time with various texture patterns and appearances. In generating the texture patterns of leaves, different from the existing technique that models the appearance manifolds from a single input material sample, this paper captures the BRDF and BTDF properties of leaf surface from multiple various samples, which contribute to a final complete aging manifold representing the appearance change through the aging process of the leaves. Combining this aging manifold representing with botanical knowledge, it can synthesize many different leaves' appearance and textures and extrapolate texture patterns out of input samples. The effectiveness of this method is proved in several applications, including subjects with different ages and types. In most cases, it is able to re-produce texture patterns consistent with those observed in nature. © by Institute of Software, the Chinese Academy of Sciences. All rights reserved.