During the laser powder bed fusion (LPBF) additive manufacturing (AM) process, the printed part dimension accuracy, surface quality, and mechanical properties highly depends on the process stability, including the metallic ejected spatters formation and melt pool geometry. During the LPBF process, a high vaporization recoil force generated on the liquid metallic material within the melt pool, along with Marangoni force, surface tension force, and buoyance force, result in unpredictable and undesirable metallic ejected spatters and unstable melt pool behavior. Both the metallic ejected spatters and unstable melt pool play a significant influence on the deposited part’s surface roughness, dimension accuracy, defects formation, and mechanical properties. To investigate the process stability of LPBF AM process and print high-quality AM parts, we apply a high-speed optical camera along with EOS M290 to capture the complex spatters formation and behavior, melt pool size and geometry change during the LPBF process. A large volume of images is recorded during the LPBF process, which makes it difficult to manual analysis. Here, to analyze the LPBF process stability under different process conditions, we apply an effective image analysis method using home-designed Python code to identify the spatters’ generation and motion and detect melt pools size and geometry using the recorded images. This image analysis algorithm can accurately detect the generated spatter amount, size, and its initial ejection speed. In addition, this algorithm can identify the newly generated spatters and previous existing spatters with a high accuracy. Moreover, we can extract the information of the moving melt pool’s size and stability using this image analysis method. The LPBF process stability of videos recorded under two different process conditions are evaluated by this image analysis algorithm. The spatter amount, size, and initial ejection speed are detected and analyzed. In addition, the moving melt pool’s size and stability are estimated using this algorithm. Based on the extracted information of spatter formation and behavior, and melt pool size and stability, we propose a stability index to estimate the LPBF process stability. A high stability index indicates more spatters formation and large variation of melt pools size and during the LPBF process. In other words, a small stability index indicates small number of spatters and more stable melt pool geometry, may contributing to produce high quality AM parts with smooth surfaces and high dimension accuracy. The proposed image analysis algorithm and the stability index are helpful to evaluate the LPBF process behavior according to the extracted information of spatters and melt pool. Similar image analysis algorithm and stability index can also be applied to estimate the process stability and robustness for other manufacturing processes using different alloys.