Abstarct: Technical analysis-baxsed strategies face challenges such as optimal parameter selection and the risk of overfitting, due to their strong dependence on input parameters like the lookback period. The manual selection of these parameters is often a subjective, trial-and-error process that can lead to poor strategy performance in changing market conditions. This research addresses the existing gap in this field by presenting a systematic approach for the automated optimization of key parameters of a trading strategy using mextaheuristic algorithms. In this study, a trading strategy baxsed on identifying peaks and troughs in the Exponential Moving Average (EMA) of asset prices is employed. To optimize the key parameter of this strategy, namely the EMA period, the Grey Wolf Optimizer (GWO) algorithm was used. Inspired by the social behavior of wolves, this algorithm searches the parameter space to find a value that maximizes portfolio value. The model was trained on historical data of commodity futures (including gold, silver, and crude oil) from the beginning of 2020 to 2024, and its performance was evaluated on out-of-sample data from 2024. The results showed that the GWO-optimized strategy generated significant returns during the in-sample training period. The evaluation of the strategy on out-of-sample (test) data also demonstrated positive performance, leading to an average profitability of 85.78% in the asset portfolio. However, this return was accompanied by higher volatility, which highlights the challenges of model generalization in real markets. This research confirms the high potential of mextaheuristic algorithms in automating and improving trading decisions baxsed on technical analysis and provides a data-driven frxamework for developing trading strategies.