Textures can be often observed in images, motions, and surface geometries, and thus they can be used to describe a wide range of natural phenomena which vary randomly over repeat patterns. Directed to the major goal of realism for expressing natural world, textures play a significant role in rendering synthetic images and animations. This thesis focuses on texture synthesis and analysis techniques, and a number of applications related to textures and images. This thesis firstly proposes a novel iterated random search model for large-scale texture synthesis and manipulations. This approach is motivated by the theory of Markov Random Field (MRF) which drives our bold guess that a random sampling may just catch a good match and can propagate the natural coherence in the neighborhood. On the other hand, we use an iteration process to constantly update certain bad guesses so that the algorithm can converge fast and generate improved results in the state of art. Previous researches on texture synthesis and manipulations have generated good results both on quality and performance, but their cost remains high especially for large-scale and complex textures. Our algorithm achieves one or two orders of magnitude faster than the typical example-based texture synthesis methods. A simple theoretical analysis is also provided to compare our model and the classical synthesis algorithms. Besides, this simple method turns out to work well in various applications such as texture transfer, image completion and video synthesis. Another work presented in this thesis is an image completion and editing task in perspective scenes which are demonstrated largely in the real world but rarely studied in the field of image completion. However, considering that it is quit hard to extract perspective information from a single image, this work starts from a friendly and portable interactive platform to obtain the basic perspective data. Then, in order to make this interface less sensitive, easier and more flexible, a perspective-rectificationbased correction mechanism is proposed to iteratively update the locations of the initial points selected by users. Finally, we have also explored an automatic algorithm after the extensive investigation on architectures which often contain a lot of repetitions and symmetries. Our experiments show the results with good qualities and illustrate the validity of our approaches by various images and applications.