In the commercial sector, a newly formed enterprise is directed by an entrepreneur to make and market new services, and products. The identification of hopeful companies is required by investors and creditors. Hence, the prediction of startup survival evaluates the potential capability and development of a new commercial venture. Hence, this paper proposes the Eel and Grouper Weighted Moving Average-based Spinal Zeiler and Fergus Network (EGWMA_SpinalZFNet-based startup success prediction. The input data is preprocessed with the utilization of Z-score normalization, in which the noise from the data is eliminated. The required features are selected using the City Block Distance. The SpinalZFNet predicts the startup success, and the parameters of SpinalZFNet are trained by the proposed EGWMA. Moreover, the EGWMA_SpinalZFNet-based startup success prediction is estimated using the accuracy, sensitivity, and specificity metrics, in which the optimal values of 91.40%, 92.22%, and 90.38% are attained.