Batteries have been the focus of attention due to their numerous advantages in distinct applications such as recently on Electric Vehicles A limiting factor for adaptation by industry is related to the aging of batteries over time. Characteristics of battery aging vary depending on many factors such as battery type, electrochemical reactions and operation conditions. Aging could be considered in two sections according to its type: calendaring and cycling. This thesis presents a review of empirical and semi-empirical modelling techniques and studies of aging. It focuses on the trends observed across different studies for two types of aging and highlights the limitations and challenges of various models. It introduces three different models for semi-empirical modelling based on Arrhenius Law from the literature for calendar aging, which cover all three important factors for calendar aging: temperature, stage of charge, and time. Moreover, four more models are developed based on these three factors and the Arrhenius law to contribute to the literature. To examine the usability and compatibility of these models, we selected five different experimental sets based on different chemistries and operating conditions from the literature. We also added calendar aging experiments carried out within the scope of our HORIZON-Helios European Project and examined a total of six experimental sets. The Helios Project dataset is split into 70% training data and 30% prediction data to measure the ability to predict future capacity loss. For this purpose, linear regression and genetic algorithm methods were used to determine the parameters of each semi-experimental model by minimizing the mean square error value between the prediction results and experimental capacity data. As a result, it was seen that the numerical solution obtained using the genetic algorithm gave better results than the analytical solution obtained by linear regression. The objective of this thesis is to present comprehensive and accurate models by examining the compatibility of models proposed in the literature models developed in our research with experimental sets. 7 Semi-Empirical Models (SEM), based on a fixed set of defined parameters, have obtained satisfactory estimates of calendar obsolescence for given storage conditions. SEM-3 and 7 were able to predict capacity loss with low errors. In particular, SEM-3 had the lowest RMSE in most experimental sets. While model errors are generally close to each other, Redondo-Iglesias et. al model and Model 7 have lower errors, similar to SEM-3. When all data sets are examined, it is seen that the lowest and highest RMSE values in the model predictions are 0.036 and 3.91, respectively.