The thermochemical conversion of biomass into producer gas presents an attractive alternative fuel option for compression ignition (CI) engines, making biomass gasification a critical driver for achieving sustainable development goals. Considering the application of producer gas (PG) in CI engine, the most potential gases include H2 and CO as main fuel compounds and it is crucial to comprehensively understand the impact of these two gas components on the engine behaviour. Nowadays, artificial intelligence-powered models are frequently applied for simulating engines that run on a single type of fuel. However, their usage is not as common when it comes to modeling dual-fuel CI engines run on synthetic producer gas or syngas. The present study explores the feasibility of optimizing operational parameters, such as engine load and syngas composition, in improving the efficiency and lowering the levels of pollutants emitted by a 3.5 kW CI engine operated under dual fuel (DF) mode using syngas as primary fuel and diesel as pilot fuel. The performance and emission characteristics of syngas (H2:CO) is examined by studying its behaviour in four different combinations. The compositions of syngas are prepared based on the volumetric percentage of the H2 and CO and is inducted into the combustion chamber using a novel venturi-type air-gas mixer. In the present study, an intelligent metaheuristics-based optimization algorithm i.e., Teaching–Learning Based Optimization (TLBO) is developed and introduced, to develop a predictive model within constrained range of engine operating conditions. Further, the algorithm is used to estimate multiple engine performance characteristics simultaneously viz., brake thermal efficiency (BTE), unburned hydrocarbons (HC), and carbon monoxide (CO). The resultant findings identify the optimal engine load of 68.87% and the ideal syngas composition of 63.9% H2 and 49.5% CO as key parameters for maximizing engine efficiency while minimizing exhaust emission. At these optimized operating condition, 19.49% BTE is observed, while HC and CO emission was found to be 384.6 ppm and 445.33 ppm respectively. This shows the effective and efficiency of the proposed algorithm.