Abstract: Plantation carbon emissions are the significant sources of greenhouse gas emissions. Accurate prediction and effective management of these emissions are crucial for mitigating climate change and promoting sustainable agricultural development. Conventional prediction models exhibit limited capability in capturing the complex nonlinear interactions inherent in the plantation carbon emission system, and their insufficient robustness often leads to overfitting. This study takes the plantations in Heilongjiang Province as a case study to explore how to optimize existing methods for predicting plantation carbon emissions. First, the IPCC carbon emission method is applied to comprehensively account for three major sources of carbon emissions: agricultural land use, CH₄ emissions from rice fields, and N₂O emissions from agricultural land. The carbon emissions from planting activities in Heilongjiang Province from 2001 to 2022 were systematically calculated. Based on this, a long short-term memory (LSTM) network model was developed, incorporating three key dimensions: social and economic drivers, production scale effects, and technical energy consumption intensity. To enhance the model’s predictive performance, the improved whale optimization algorithm (IWOA) was introduced to optimize four hyperparameters of the LSTM model: the number of hidden units, learning rate, batch size, and training epochs. Finally, the IWOA-LSTM model was used to predict future plantation carbon emissions in Heilongjiang Province from 2023 to 2027 under both baseline and low-carbon scenarios. The results indicate several findings: 1) The plantation carbon emissions in Heilongjiang Province show a trend of “rapid growth followed by fluctuating decline”, reaching a peak of 20.45 million tons in 2015. The main sources of carbon emissions include CH₄ emissions from rice fields, N₂O emissions from agricultural land, and carbon emissions resulting from fertilizer production and application, and their average proportions in the total annual emissions are 41.42%, 38.26% and 11.65%, respectivley. 2) Compared to the unoptimized LSTM model, the IWOA-LSTM model demonstrates significant improvements in both prediction accuracy and stability. It achieves a mean absolute error of 55.82×10⁴ t, a root mean square error of 61.74×10⁴ t, and a mean absolute percentage error of 2.83%, all of which are superior to the LSTM model’s 114.41×10⁴ t, 124.72×10⁴ t, and 5.78%, respectively. The study demonstrates that the IWOA-LSTM model can effectively predict plantation carbon emissions, providing a scientific basis for the formulation of carbon reduction policies for Heilongjiang Province’s planting industry. 3) The prediction results of the IWOA-LSTM model show that the plantation carbon emissions in Heilongjiang Province could be effectively suppressed by controlling crop planting area, improving fertilizer application efficiency, and reducing the diesel consumption per unit area of agricultural machinery. Based on the conclusions drawn above, the following recommendations for emission reduction are proposed: optimizing land use structure and controlling crop planting area, increasing the application and innovation of green agricultural technologies, promoting rural economic development, and increasing farmers’ income, strengthening policy support, and incentive mechanisms. Through the above measures, sustainable development of Heilongjiang Province’s agriculture could be further achieved.