ARIMA models a series using its own past values (autoregressive part AR, order p), differencing (I, order d), and past errors (moving average MA, order q). SARIMA adds a seasonal component (P, D, Q, period). The orders are chosen with the ACF and PACF or criteria such as the AIC.
Example
SARIMAX(serie, order=(1, 1, 1), seasonal_order=(1, 1, 1, 12)).fit()