Body weight is an important indicator for reflecting the health and growth conditions, reproduction and production performance of livestock and poultry. Accurate and rapid assessment and monitoring of livestock and poultry body weight is a critical way to improve the level of breeding management and achieve precision livestock farming. The traditional weighing method is time-consuming and laborious, and easy to cause stress response on animals. Weight assessment based on machine vision technology, which can establish an intelligent assessment model between body weight and body shape characteristics by using visual detection technology, is a hotspot of intelligent technology research in livestock and poultry breeding at present. Firstly, the methods of weight assessment were categorically described. Then, the sensor types, methods and applications of animal and poultry body feature treatment were analyzed in detail. The comparative analysis of the research on body size, physical signs and weight assessment model based on machine learning method were focused on. The application effect and the latest research results of various machine learning algorithms in weight assessment were presented. The development potential of deep learning algorithm in the field of automatic weight assessment of livestock and poultry was discussed and analyzed. Finally, the problems and challenges of weight assessment researches on livestock and poultry and the development trend of the future work were pointed out, which can provide some references for the scholars and engineers in the field of the modern intelligent weight assessment for livestock and poultry.