To enable next generation battery chemistries with higher energy density or expand the application of current lithium-ion batteries (LIBs), novel electrolytes are needed to replace the commercially used carbonate-based electrolytes. In this dissertation, a molecular design strategy is proposed and tested in several electrolyte solvent classes including fluorinated ethers, nonfluorinated ethers and borate esters. The effects of solvent molecular structure on the physicochemical and electrochemical properties of the electrolytes are investigated systematically. Through rational molecular design, the electrolyte performance in batteries can be tuned and optimized. In Chapter 2, the molecular design of fluorinated ether solvents is explored for lithium metal batteries (LMBs). By changing the ratio between fluorinated group and ether group, the ion solvation structure of fluorinated ether electrolytes is tuned. Strong ion pairing is found to be beneficial for lithium metal compatibility but leads to poorer oxidative stability. Chapter 3 highlights a surprising discovery that fluorinated ether electrolytes are compatible with graphitic anodes, which expands their application to LIBs. A set of electrochemical and physical characterizations prove that fluorinated ether solvents can suppress solvent co-intercalation challenges that commonly plague nonfluorinated ether solvents at moderate salt concentrations. Compared to commercial carbonate electrolytes, a presentative fluorinated ether electrolyte, E3F1, enables better thermal stability and improved compatibility with silicon containing anodes. In Chapter 4, the molecular design strategy is applied to borate ester solvents and the effect of fluorination degree on solvation ability is studied. Moderate fluorination using -CH₂F group was shown to be beneficial for ion solvation as the moderately fluorinated borate ester TFEB shows higher lithium salt solubility than its nonfluorinated or perfluorinated counterparts. Ab initio molecular dynamic simulation indicates that TFEB solvent utilizes -CH₂F group as the primary coordination site to lithium ion. In Chapter 5, the molecular structure of fluorinated ether solvents is further optimized for fast charging and all temperature LIBs. The -CF₃ group in E3F1 is replaced with -CH₂F group and ether backbone length is decreased to lower solvent molecular weight for faster ion transport. The optimized fluorinated ether electrolyte (E1CH2F) enables fast charging and wide working temperature range in 4.4 V class LIBs. In Chapter 6, the molecular design of nonfluorinated ether solvent is investigated for anode free LMBs. A group of sterically hindered ether solvents (xMEs) are designed by introducing longer alkyl end groups on the backbone of 1,2-dimethoxyethane (DME). The weaker solvation power promotes more ion pairing in xME electrolytes than DME, which leads to improved anode free LMB cycle life and better oxidative stability. In Chapter 7, exploratory attempts are made on using an active learning (AL) approach to accelerate electrolyte solvent discovery for anode free LMBs. A Gaussian process regression (GPR) surrogate model is trained initially with an in-house anode free LMB cycling data set and then applied on a pre-defined search space with cell capacity retention as the target property to optimize. After six active learning campaigns, the model was able to filter out seven compounds with promising cycling behavior. Four of the seven electrolytes show excellent performance that is comparable to state-of-the-art electrolytes in the literature. Lastly, a conclusion chapter summarizes the molecular design principles and discusses some promising future directions for electrolyte development for next generation batteries.