Soft robotics, a recent advancement in robotics systems, distinguishes itself by utilizing
soft and flexible materials like silicon rubber, prioritizing safety during human interaction, and
excelling in handling complex or delicate objects. Soft pneumatic actuators, a prevalent type of
soft robot, are the focus of this paper. A new geometrical parameter for soft artificial pneumatic
muscles is introduced, enabling the prediction of actuation behavior using analytical models based
on specific design parameters. The study investigated the impact of the chamber pitch parameter and
actuation conditions on the deformation direction and internal stress of three tested soft pneumatic
muscle (SPM) models. Simulation involved the modeling of hyperelastic materials using finite
element analysis. Additionally, an artificial neural network (ANN) was employed to predict pressure
values in three chambers at desired Cartesian positions. The trained ANN model demonstrated
exceptional performance. It achieved high accuracy with training, validation, and testing residuals
of 99.58%, 99.89%, and 99.79%, respectively. During the validation simulations and neural network
results, the maximum errors in the x, y, and z coordinates were found to be 9.3%, 7.83%, and 8.8%,
respectively. These results highlight the successful performance and efficacy of the trained ANN
model in accurately predicting pressure values for the desired positions in the soft pneumatic muscles.