One of the important assumptions of the premium pricing of a health insurance product is the probability for someone suffers from a certain disease. In this paper, the disability income model is applied to a company using two covariates, namely age and sex. The purpose is to find out the magnitude of the probability for employees to experience disabilities due to work, a multi-state model can be used with semi-Markov assuming. There are several approaches to complete the multi-state model, one of which is the transition intensity approach. The intensity of the transition in this paper is estimated using the maximum likelihood approach, which will produce a crude estimate. Afterwards, the graduation process is performed on a crude estimate to obtain a finer shape of the transition intensity function with the Generalized Linear Model (GLM). The intensity of the transition from the graduation results is used to form transition probabilities which are eventually used as one of the assumptions in premium pricing.