Environmental changes create significant threats to agricultureAgriculture, mainly animal production. These fluctuations will negatively impact the poultry farm and reduce production due to environmental changes. In order to improve the production of poultry farms, the effective management of air, water and heat stress quality is essential. These parameters are collected manually by the poultry farm, but the manual predictionPrediction is complex. For this reason, several existing methods are developed to predict air quality, water quality and heat stress automatically, but adequate predictionPrediction is not achieved. This paper's proposed method is based on Deep LearningDeep learning (DL) algorithms for automatically predicting poultry production. The proposed method calculates the Water Quality IndexWater quality index (WQI), Humidex valueHumidex value and Air Quality IndexAir quality index (AQI) using manually collected parameters like temperature, humidity, etc. The WQI, AQI and heat stress estimates are helpful to control air quality, water quality and heat stress in poultry farms and thus increase production. In the proposed method, the M-squared normalization (MSN) method is used for data preprocessing and Duo attentionDuo attention with depth-wise separable convolutional neural networkDepth-wise separable Convolutional neural network (Duo-SCNN) to predict the production of poultry farms. This method uses the combined air, water and heat parameters dataset to calculate the WQI, AQI and heat stress. The proposed method was evaluated using several performance metrics such as accuracy, precision and recall. The proposed method achieved a high accuracy of 97.5% in poultry farm predictionPrediction.