Task scheduling in distributed computing environments is a complex problem that requires careful consideration of resource allocation and system efficiency. Efficient task scheduling in cloud and fog computing is challenging due to diverse tasks, high latency, and insufficient processing times, especially in applications requiring strict computational efficiency. To address these challenges, enhancing energy efficiency in cloud workflow scheduling through mixed generative adversarial imitation learning to minimize power consumption in distributed cloud environments (CWS-MixGAIL) is proposed. Initially, input image is collected from Google 2019 Cluster Sample Dataset. Then, the input images are pre-processed using Bilinear Double-Order Filter (BDOF) is used to handling the missing values and removing duplicates. Then, the pre-processed data are fed into prediction. The mixed generative adversarial imitation learning (MixGAIL) is used to predict the energy consumption patterns in cloud workflow scheduling. By then, the proposed method is integrated into the MATLAB workspace and the proposed CWS-MixGAIL outperforms others with 140 Kbps throughput, 98.5% efficiency, 50–150 J energy consumption, and 75–80% fault tolerance, with existing methods, like an efficient deep reinforcement learning based task scheduler in cloud-fog environment (TS-CFE-CNN), efficiency-aware adaptive deep reinforcement learning for dynamic task scheduling in edge-cloud environments (EA-ECE-DRL) and cloud-edge hybrid deep learning framework for scalable IoT resource optimization (CE-RO-DQN).