This paper presents a novel term structure of interest rate (TSIR) model with stochastic volatility and jumps (SVJ) that combines the market framework proposed by Brace et al. (1997) with the string-shock framework of Santa-Clara and Sornette (2001). In this model, the factors’ variance is estimated through the eigendecomposition of a variance–covariance matrix obtained with a measure of market volatility derived from out-of-the-money option prices and historical correlations of interest rates traded in the futures market. The stochastic evolution of the factors’ variance is governed by the 4/2 model developed by Grasselli (2017), including jumps. A novel method is employed to estimate the parameters of the SVJ model that minimizes the distance between the sample moments and the moments of a gamma distribution. The empirical application of the model in the Brazilian derivatives market demonstrates its effectiveness in accurately capturing the volatility smile of interest rate options. Display omitted •Introduces a new TSIR model with stochastic volatility and jumps.•Proposes a novel methodology to estimate the time-varying variance–covariance matrix.•Proposes a novel methodology to estimate parameters of the stochastic volatility model.•The model capture the volatility smile of IDI options in the Brazilian financial market.•Analyzes the predictive power of interest rate futures and the options market.