Electric vehicles (EVs) have become the most reliable option to replace the current automotive fleet, driven by significant battery technology advances, competitive battery prices, government policies and incentives, and global warming and climate change concerns. The reliability of EVs mostly depends on the performance of lithium-ion batteries (LIBs), currently the preferred energy storage choice in EVs. The development and improvement of LIB chemistries, along with the design of effective Battery Management Systems (BMS) and Battery Thermal Management Systems (BTMS), have contributed to maximizing the lifetime of the LIBs. Despite these technological advances, degradation, commonly referred to as aging, is inevitable and irreversible, affecting battery performance over time, causing a decrease in power and capacity. Accurate battery health diagnosis and prognosis are therefore crucial for describing the aging phenomenon. Aging is mostly described and predicted as the evolution of battery performance parameters in time, through limited correlations based on accelerated test aging data or data-driven models supported by machine learning techniques. These approaches do not allow for the identification and segregation of the mechanisms responsible for aging, which would provide a valuable aging input to the BMS for optimal control actions, accurately setting key battery lifetime thresholds. On the other hand, electrochemically-based aging models are robust but computationally expensive and have not been successfully implemented in BMS algorithms. Data-based and electrochemically-based aging models have been mostly developed at the cell level. Moreover, batteries experience intrinsic cell-to-cell variations or spreading that, in addition to uneven extrinsic stress factors, cause different degradation trajectories per cell, contributing to cell imbalance. To overcome these challenges in modelling aging in LIBs, and hence to support the design and control strategies of both BMS and BTMS aiming to maximize the battery lifetime, this thesis work proposes an aging prediction framework in lithium-ion batteries by: (i) using a statistical-kinetics approach to obtain a set of test matrices and modelling guidelines for data-based models from aging accelerated tests; (ii) integrating an electrochemical reduced-order performance model with an electrochemical aging diagnostic tool, allowing for the identification and segregation of the main degradation mechanisms, and for the accurate identification of battery life thresholds such as first-, second- and unusable-life; and (iii) ensuring the scalability of the proposed approach from cell to battery level by considering cell spreading, using a stochastic approach to predict the state-of-health (SOH) of the battery.