Options and derivative products are complicated financial tools. Because of the risk involved in options trading, trade support systems are in high demand to help clients manage and mitigate their volatility. The major issue of option price forecasting is the non-linearity and non-stationarity of option characteristics. The key job in options trading is to calculate the realistic price option, which may be used to assess which options are now inexpensive and which are currently expensive. To address these issues in option price forecasting, a Two-Tier Machine Learning Ensemble model (TTMLE) using Salp Swarm Optimization (SSO) has been proposed. In the TTML model, Arbitrary Mode Disintegration (AMD) and Composition Search Algorithm (CSA) has been integrated into tier 1 and tier 2, respectively to forecast option pricing. The salp swarm optimization technique has been used to teach the TTMLE model and improve the weights of the model, resulting in the TTMLE-SSO model, which enhances forecasting accuracy. Other models such as the Non-linear Neural Network (NNN), Support Vector Regression (SVR), and Long Short-Term Memory (LSTM) network have been compared to the suggested model. The new approach beats previous methods, and predicting accuracy is substantially improved, according to empirical data.