Integrated Energy Systems (IES) play a crucial role in addressing energy supply-demand imbalances and enhancing energy utilization efficiency. This paper presents Energy Management Decision Transformer (EMDT), a novel approach aimed at improving decision-making efficiency within IES. Unlike traditional methods that rely on precise modeling, EMDT employs the causal transformer framework to streamline decision-making processes. Methodologically, we propose a conditional sequence modeling approach, leveraging pre-collected datasets to extract insights from historical data and introducing reward paradigms and training techniques to optimize IES operations. Through extensive evaluation over thousands of operational days within IES, we demonstrate the effectiveness of the EMDT system in enhancing the stability of energy utilization systems and its ability to adapt to fluctuating demands and price conditions. This research contributes to advancing the field of energy management by providing a versatile and adaptive framework for decision-making in complex energy systems.