The rise of cryptocurrencies as a new asset class has introduced significant challenges in investment management due to their extreme volatility, financial bubbles, regulatory issues, cybersecurity risks, and project failures. Given cryptocurrencies' vast diversity and rapidly changing market values, effective risk management and optimal portfolio construction are crucial. This study explores key factors in cryptocurrency portfolio formation and employs a multi-criteria approach based on the Preference Factor to select optimal assets. Portfolio optimization is conducted using four objective functions: Sharpe ratio maximization, the Markowitz mean-variance model, tracking error minimization, and Value at Risk (VaR) minimization. Six optimization methods are compared: SLSQP, Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Artificial Bee Colony (ABC), Differential Evolution (DE), and the Markowitz model. Results indicate that SLSQP performs best when maximizing the Sharpe ratio, achieving a superior risk-return balance. Statistical analyses confirm that the optimized portfolio significantly outperforms a naive benchmark. Sensitivity analysis reveals that extremely small or large initial weights lower the Sharpe ratio, while excessive iterations add computational costs without notable performance improvement. This research utilizes historical data from 183 cryptocurrencies with the highest market capitalization over three years (January 2019 – January 2024) and evaluates results using numerical and statistical methods. The findings provide valuable insights into constructing optimized cryptocurrency portfolios with robust risk management strategies.