This study focuses on the processing and analysis of force and tool holder acceleration signals in multiple directions during a turning of high-strength alloy steel (42CrMo4+QT) workpiece in high accuracy operation case, with the objective of identifying the features that are strongly correlated with surface quality. Force signals are first subjected to filtering to remove noise, after which several time-domain features such as average value, standard deviation, maximum, and minimum are extracted. In addition, acceleration signals are analyzed in the frequency domain to obtain spectral features that may reflect dynamic cutting behavior. The extracted features are statistically evaluated for their correlation with the machined surface roughness, which is used as the indicator of surface quality. The study enabling a systematic evaluation of their influence on surface roughness under varying cutting conditions. The results show that the corner radius, feed, and depth of cut are important cutting parameters, while average cutting force features also shows better correlation compared with acceleration features. The study demonstrates that these features have relatively adequate R2 and p-values within the ideal limit. These feature can serve as reliable indicators for predicting surface quality, potentially enabling real-time monitoring and quality control in the high strength alloy steel turning operations.