<![CDATA[Abstract. Stock price is an important indicator reflecting a company's performance and serves as a basis for investment decision-making. The fluctuating movement of stock prices requires a forecasting method capable of producing accurate short-term predictions. This study aims to determine the best Autoregressive Integrated Moving Average (ARIMA) model for forecasting the stock price of PT Bank Rakyat Indonesia (Persero) Tbk (BBRI). The data used consist of daily closing stock prices from January 1, 2025 to August 15, 2025, totaling 143 observations. The analysis stages include stationarity testing using the Box-Cox transformation and the Augmented Dickey-Fuller (ADF) test, model identification through the Autocorrelation Function (ACF) and Partial Autocorrelation Function (PACF), parameter estimation using Maximum Likelihood Estimation (MLE), model evaluation based on Residual Mean Square (MS) and the Ljung-Box diagnostic test, and forecasting accuracy measurement using Mean Absolute Percentage Error (MAPE). The results indicate that the ARIMA(0,1,2) model is the best forecasting model with the smallest Residual Mean Square (MS) value of 10,132.8 and residuals satisfying the white noise assumption. Forecasting results show that the stock price of PT Bank Rakyat Indonesia (Persero) Tbk tends to remain stable over the next seven periods with excellent forecasting accuracy based on the MAPE value. Keywords: ARIMA, forecasting, stock price, time series, BBRI. Abstrak. Harga saham merupakan salah satu indikator penting yang mencerminkan kinerja perusahaan serta menjadi dasar pengambilan keputusan investasi. Pergerakan harga saham yang bersifat fluktuatif menyebabkan diperlukan metode peramalan yang mampu menghasilkan prediksi jangka pendek secara akurat. Penelitian ini bertujuan untuk menentukan model Autoregressive Integrated Moving Average (ARIMA) terbaik dalam meramalkan harga saham PT Bank Rakyat Indonesia (Persero) Tbk (BBRI). Data yang digunakan merupakan data harga penutupan harian saham BBRI periode 1 Januari 2025 sampai dengan 15 Agustus 2025 sebanyak 143 observasi. Tahapan analisis meliputi uji stasioneritas menggunakan transformasi Box-Cox dan uji Augmented Dickey-Fuller (ADF), identifikasi model melalui plot Autocorrelation Function (ACF) dan Partial Autocorrelation Function (PACF), estimasi parameter menggunakan Maximum Likelihood Estimation (MLE), evaluasi model berdasarkan Residual Mean Square (MS) dan uji diagnostik Ljung-Box, serta pengukuran akurasi menggunakan Mean Absolute Percentage Error (MAPE). Hasil penelitian menunjukkan bahwa model ARIMA(0,1,2) merupakan model terbaik dengan nilai Residual Mean Square (MS) sebesar 10.132,8 serta memenuhi asumsi residual white noise. Hasil peramalan menunjukkan harga saham PT Bank Rakyat Indonesia (Persero) Tbk cenderung stabil pada tujuh periode mendatang dengan tingkat akurasi peramalan yang sangat baik berdasarkan nilai MAPE. Kata Kunci: ARIMA, harga saham, peramalan, time series, BBRI.]]>
