Background: Several instruments have been developed to measure nursing workload. The commonly used Nursing Activities Score (NAS) and Therapeutic Intervention Scoring System (TISS) are applied to all types of ICU patients. Former research showed that NAS explained 59 to 81% of actual nursing time, whereas the Therapeutic Intervention Scoring System (TISS) described only 43% of the actual nursing time. In both models the development was not based on time measurements. Objectives: The aim of this study was to develop a time-based model which can assess patient related nursing workload more accurately and to evaluate whether patient characteristics influence nursing time and therefore should be included in the model. Design: Observational study design. Setting: All 82 Dutch ICUs participate in the National Intensive Care Evaluation (NICE) quality registry. Fifteen of these ICUs are participating in the newly implemented voluntary nursing capacity module. Seven of these ICUs voluntarily participated in this study. Participants: The patient(s) that were under the responsibility of a chosen nurse were followed by the observer during the entire shift. Methods: Time spent per nursing activity per patient was measured in different shifts in seven Dutch ICUs. Nursing activities were measured using an in-house developed web application. Three different models of varying complexity (1. nursing activities only; 2. nursing activities and case-mix correction; 3. complex model with case-mix correction per nursing activity) were developed to explain the total amount of nursing time per patient. The performance of the three models was assessed in 1000 bootstrap samples using the squared Pearson correlation coefficient (R2), Root Mean Squared Prediction Error (RMSPE), Mean Absolute Prediction Error (MAPE), and prediction bias. Results: In total 287 unique patients have been observed in 371 shifts. Model one's Pearson's R was 0.89 (95%CI 0.86–0.92), model two with case-mix correction 0.90 (95%CI 0.88–0.93), and the third complex model 0.64 (95%CI 0.56–0.72) compared with the actual patient related nursing workload. Conclusion: The newly developed Nurse Operation Workload (NOW) model outperforms existing models in measuring nursing workload, while it includes a lower number of activities and therewith lowers the registration burden. Case-mix correction does not further improve the performance of this model. The patient related nursing workload measured by the NOW gives insight in the actual nursing time needed by patients and can therefore be used to evaluate the average workload per patient per nurse.