Workers' poor mental status (e.g., stress and mental fatigue) is a critical factor in accidents, errors, and loss of productivity at construction sites. With recent advances in wearable technologies, workers' mental status can be monitored by examining patterns of brainwaves in their electroencephalogram (EEG) signal. Acquiring high-quality EEG in the field using a wearable device is a critical step in assessing worker mental status because there are several types of significant signal artifacts (i.e., noise) can interfere with the actual signal. Among them, ocular signal artifacts (e.g., eye movements and blinks) are the most challenging because they share a similar frequency range with EEG. To reduce ocular artifacts, researchers applied principal component analysis (PCA) and independent component analysis (ICA) method to separate intrinsic artifacts from workers' brainwave signals. Though it reduces ocular signal artifacts, it is limited to detect and reduce dependent signal noises, particularly ocular artifacts (e.g., eye blinking and eye movement). To address this limitation, the authors apply a dependence component analysis (DCA) method for detecting correlative signal artifacts. Nine construction workers' brainwaves collected at multiple construction sites were used to examine the performance of the proposed method. Then, the performance of the DCA is compared with two traditional methods, PCA and ICA. According to the results, there are significant improvements in the output signal-to-noise ratio (SNR) and mean square error (MSE) compared to ICA and PCA. The results confirm that the proposed DCA method outperforms the traditional methods in reducing ocular signal artifacts from a wearable EEG headset at construction sites. The result opens up the opportunity to capture a higher quality EEG and get a better assessment of workers' mental status in the field using a wearable EEG in the field.