To address the emerging market of digital currencies where researchers have less applied deep reinforcement learning, this paper proposed deep evolutionary reinforcement learning (DERL) model, which is a deep reinforcement learning model based on an evolutionary strategy to extract trading information between users using Bitcoin K-line data as base data. Reinforcement learning is applied to data cleaning and factor extraction from a high-frequency, microscopic viewpoint to quantitatively explain the supply and demand imbalance and to create trading strategies. In order to determine whether the algorithm can successfully extract the significant hidden features in the factors when faced with large and complex high-frequency factors, this paper trains the agent in reinforcement learning using three different learning algorithms, including Q-learning, evolution strategy, and Policy Gradient. The experimental dataset, which contains data on sharp up, sharp down, and continuous oscillation situations, was chosen to test Bitcoin in January-February, September, and November 2022. According to the experimental results, the evolution strategy algorithm achieved returns of 59.18%, 25.14%, and 22.72%, respectively. The results also demonstrate that deep reinforcement learning based on the evolution strategy outperforms Q-learning and Policy Gradient concerning risk resistance and return capability.