Freshwater scarcity remains a critical global challenge, particularly in arid and semi-arid regions, driving growing interest in decentralized and small-scale desalination technologies. Humidification-dehumidification (HDH) systems assisted by thermoelectric cooling (TEC) offer compact, controllable dehumidification; however, their practical deployment is inherently constrained by the trade-off between freshwater productivity and energy efficiency. In this work, a laboratory-scale HDH-TEC desalination system incorporating an evaporative pad humidifier and a thermoelectric dehumidifier was experimentally investigated under ambient temperatures ranging from 28 to 43 degrees C, air velocities from 1.5 to 2.5 m/s, and TEC voltages from 6 to 12 V. Artificial neural networks were employed as surrogate models to predict freshwater condensation rate and system effectiveness based on experimentally validated operating conditions. Among four tested Artificial Neural Network (ANN) architectures, the configuration with 8 and 16 neurons in the two hidden layers provided the most reliable performance, achieving coefficients of determination (R2) above 0.99 for both output parameters. A decisionoriented multi-objective assessment based on weighted scoring was then introduced to quantitatively evaluate the competing objectives of water productivity and energy efficiency under different operational priorities. The experimental results demonstrate that increasing TEC voltage significantly enhances freshwater production, exceeding 40 mL/h at low ambient temperature, while simultaneously reducing the coefficient of performance (COP) and overall system effectiveness. The multi-objective analysis identifies moderate TEC voltages between 8 and 10 V combined with an air velocity of 2.0 m/s as a balanced operating regime under mild climatic conditions, providing practical guidelines for the operation of thermoelectric-assisted HDH desalination systems.