Liquid desiccant air conditioning (LDAC) has received much attention in recent years. This is mainly because LDAC systems are able to control latent loads in a more energy efficient way than conventional air conditioning systems. Although many research studies have been conducted on LDAC technologies, the following gaps in the scientific literature are addressed in this thesis: (1) carryover of desiccant droplets in air streams, (2) direct comparisons between different configurations of LDAC systems, (3) fundamentals of capacity matching in heat-pump LDAC systems, (4) optimal-control strategies for heat-pump LDAC systems, and (5) importance of transients in evaluating the performance of a LDAC system. Items (1) to (4) are addressed using TRNSYS simulations, and item (5) is addressed using data collected from a field test. The use of liquid-to-air membrane energy exchangers (LAMEEs) as dehumidifiers and regenerators in LDAC systems eliminate the desiccant droplets carryover problem in air streams. This is because LAMEE separate the air and solution streams using semi-permeable membranes, which allow the transfer of heat and moisture but do not allow the transfer of the liquid desiccant. A preliminary configuration for a membrane LDAC system, which uses LAMEEs as the dehumidifier and regenerator, is proposed and investigated under fixed operating conditions in this thesis. The influences of key design and operating parameters on the heat and mass transfer performances of the membrane LDAC system are evaluated. Results show that the membrane LDAC technology is able to effectively remove latent loads in applications that the humidity to be controlled. A comprehensive evaluation is conducted in this thesis for the thermal, economic and environmental performances of several configurations of membrane LDAC systems. The solution cooling load is covered using a cooling heat pump in all systems studied, while the solution heating load is covered using one of the following five different heating systems: (1) a gas boiler, (2) a heating heat pump, (3) a solar thermal system with gas boiler backup, (4) a solar thermal system with heat pump backup, and (5) the condenser of the solution cooling heating pump. Each of the membrane LDAC systems studied is evaluated with/without an energy recovery ventilator (ERV) installed in the air handling system. The influence of operating the ERV under balanced/unbalanced operating conditions is studied. It is found that the most economic membrane LDAC system is the one which uses the evaporator and condenser of the same heat pump to cover the solution cooling and heating loads, respectively (i.e. heat-pump membrane LDAC system). No clear guidance was found in the literature for sizing the evaporator and condenser in a heat-pump LDAC system to simultaneously meet the solution cooling and heating loads. When the heating and cooling provided by the heat pump exactly match the heating and cooling requirements of the solution, the system is “capacity matched”. A parametric study is conducted on a heat-pump membrane LDAC system to identify the influence of key operating and design parameters on achieving capacity matching. It is concluded that the solution inlet temperatures to the dehumidifier and regenerator are the most influential parameters on the moisture removal rate, capacity matching and coefficient of performance (COP). Three control strategies are developed for heat-pump membrane LDAC systems, where these strategies meet the latent loads and achieve one of the following three objectives: (1) meet the sensible loads, (2) achieve capacity matching, or (3) optimize the COP. Results show that the COP of a heat-pump LDAC system can be doubled by selecting the right combination of solution inlet temperatures to the regenerator and dehumidifier. The importance of transients in evaluating the performance of a LDAC system is addressed in the thesis using a data collected from a field test on a solar LDAC system. It is found that the sensible, latent and total cooling energy, and the total primary energy consumption of the LDAC system are changed by less than 10% during an entire test day when transients are considered. Thus, it can be concluded that steady-state models are reliable to evaluate the energy performances of LDAC systems.