On this earth, it is essential to recognize and extend compassion towards our coexisting, voiceless companions—the animals. Animal emotion detection can be a great way towards their welfare. It can involve many methods such as analysing their vocalizations and understanding their posture and physiological measures (like monitoring the heart rate and respiration). Animal emotion identification using machine learning methods is a new area of study and has garnered an increasing interest in the last few years. This work aims to identify the research challenges and gaps in this domain. In this paper, we delve into the strategies employed to detect and identify various emotions in animals using their vocalizations, posture, and their surroundings. The paper also discusses the datasets used, and the way the choice of classifiers transitioned over the years, from using machine learning models to deep learning models, thereby highlighting existing research gaps and challenges.